VIOLENCE AND LANGUAGE: The signs that hurt

By J.L. LEMKE

CAN SYMBOLIC ACTION do literal violence? Can speech cause bodily pain? Are we physically safe from hate speech and cruelty of lovers behind Cartesian firewalls separating mind from body, meaning from matter? Or do emotions breach those barriers and make us as vulnerable to words as to fists? If hypnotic suggestion can produce burn reactions, amputees feel phantom pains in long-missing limbs, and stress impairs the immune system, why should we doubt that pain is as much our response to the meaning of events as to their physical force?

In my own work,(1) these questions have emerged from an attempt to reunify our views of the symbolic and material dimensions of social processes. At Columbia, scholars interested in such diverse subjects as the sex industry, ethnic violence in South Africa, and the law of "hate speech" also find themselves trying to make sense of the effects of symbolic violence.

Violence is not simply material force: It is the use of force as a tool for some human purpose, individual or social. We are social actors and we are bodies vulnerable to pain. Every society exploits the possibility that our actions can be controlled by the fact, memory, and anticipation of pain inflicted by others. We hurt children to make them behave--sometimes with blows and sometimes with words, but equally with pain. Theories of child development make it easy to forget how often parents make children cry and how basic this violence is to the socializing process. Theories of economic and ideological domination, likewise, can obscure how the powerful exploit the powerless through pain. Violence exerts its social effects as much through what it means as through what it physically does.

Anne McClintock of Columbia's English and Comparative Literature Department has explored the transformative re-enactment in sexual relations of many basic power relationships (between parents and children, teachers and students, police and citizens, men and women).(2) In each case, the pain of social control is transmuted into sexual pleasure, and one's internalization of that control is suspended by giving another person extraordinary power over one's body in sadomasochistic rituals. Well-socialized men of high status and substantial social power reverse their roles, wearing diapers, being beaten for trivial offenses, putting themselves in the power of mothers, teachers, judges, and dominant women. In these scenes, the sight and sound of the whip is as important as its touch. McClintock vividly displays what historian Michel Foucault wrote of in Discipline and Punish: the social regulation of bodies as the foundation of the modern social order.

We demonize the extremes of violence: spousal and child abuse, gay-bashing, police brutality, prison rape, hate crimes, "ethnic cleansing." But each of these exaggerates an endemic process of violence that maintains the social order. Most violence is not idiosyncratic: The same kinds of people do the same kinds of violence to the same kinds of people. A little violence goes a long way when it takes on a meaning, when people begin to predict what will be punished. That meaning enables violence to function as a means of control. No social order could maintain itself solely by the physical effects of

violence. Violence is always also a warning, a threat of the possibility of more violence. Violence itself is a language we all learn to interpret.

Rob Nixon, also of the English department, describes the symbolic language of ethnic and class violence, particularly in the struggle between the Inkatha party and the African National Congress.(3) The axes and spears of "traditional Zulu manhood" were carried in marches and demonstrations by Inkatha, but they were more than merely symbolic in clashes with the disarmed ANC.

How can words and symbolic actions cause pain? Pain is an active response to its causes, not simply a passive effect. The meanings of words and deeds always include the feelings produced when one makes sense of a situation. The tasteless sexual or racial joke is no joke to a woman who has been raped, a man who was a victim of a hate crime, or to any person who feels less safe from pain than someone who can afford to make these kinds of jokes. We defend our freedom of speech and action, but we cannot exercise these freedoms responsibly, or judge whether others do so, if we cannot feel what hurts whom and how much.

Legal scholars like Columbia Law School's Kent Greenawalt are trying to find workable legal principles to apply to hate-speech cases. In Fighting Words: Individuals, Communities, and Liberties of Speech,(4) Greenawalt tries to balance the social interest in protecting free speech with the social and personal harms speech can do. The civil law of torts recognizes the emotional distress of harassment victims, but courts have been reluctant to accept the criminalization of speech on the basis of the damage it can do to its victims. Greenawalt cites the arguments of Mari Matsuda that "victims of vicious hate propaganda have experienced physiological symptoms and emotional distress ranging from fear in the gut, rapid pulse rate and difficulty in breathing, nightmares, post-traumatic stress disorder, hypertension, psychosis, and suicide."(5) The law of hate speech, however, is still based more on the violence it may provoke than on the equally real bodily harm it can produce.

The violence of blows and the violence of symbolic acts are not so easily separated. As sociologist Pierre Bourdieu argues in Language and Symbolic Power, people's dispositions, from the accents of their dialects to their reactions to symbolic forms, reflect the physical embodiment of their experiences. If we are what we feel, research on symbolic violence takes on special importance for the individuals and the society that this violence shapes.

Lemke, J.L. Textual Politics: Discourse and Social Dynamics. Washington, D.C.: Taylor & Francis, 1995.

McClintock, Anne. "Maid to Order: Commercial Fetishism and Gender Power." Social Text, Winter 1993.

Nixon, Rob. Homelands, Harlem, and Hollywood: South African Culture and the World Beyond. New York: Routledge, 1994.

Greenawalt, Kent. Fighting Words: Individuals, Communities, and Liberties of Speech. Princeton University Press, 1995.

Matsuda, M. "Public Response to Racist Speech: Considering the Victim's Story." 87 Michigan Law Review 2320-2336, 1989.

Resources for Attitudinal Meaning: Evaluative Orientations in Text Semantics

J L Lemke, City University of New York, Brooklyn College School of Education, Brooklyn, NY 11210 USA

1.0 Introduction

One of the most basic functions of language is to create interpersonal relationships between speakers and addressees through the way in which text is worded. Speech act functions establish whether we are offering or demanding, aiding or attacking, creating solidarity or emphasizing social distance. In these and other ways we use language to take a stance towards and socially orient ourselves and our text to others. But we do not just use language to orient to addressees, real and potential; we also take a stance toward the ideational or propositional content of our own texts. Whatever we have to say about the world, we can also tell others, in the same utterance, to what extent we believe what we say is likely, desirable, important, permissible, surprising, serious, or comprehensible. In making these evaluations of propositions and proposals, we also orient our text in the larger world of available social viewpoints on our topic, and we further define our identities as meaning-makers with particular values as well as beliefs.

Because language is a resource for doing all these important things, we need to better understand just what evaluative and attitudinal meanings it allows us to make about propositions and proposals, actions and events, persons, phenomena, and things. We want to know in what different ways the same basic sorts of attitudinal meanings can be made, and how these are instantiated in primary units such as the clause or the nominal group and across longer stretches of cohesive text. In this paper I want to identify seven semantic classes of evaluative attributes for propositions and proposals, which appear to be the only ones allowed in English. I will illustrate their use from a corpus of newspaper editorials and characterize some interesting discourse phenomena for further study.

2.0 Background

Mikhail Bakhtin (1935), a keen analyst of discourse phenomena, called attention to the ways in which social communities embody a large number of different discourse voices, each characteristic of some sub-community of speakers, and each standing in complex semantic relationships to many others. I have argued elsewhere (Lemke, 1995a) that such 'discourse formations' (cf. 'Discourses' in Gee 1990), in their mutual relationships in semantic space, mirror and help to constitute the sociological relationships among distinct groups in a community. Bakhtin (1935: 288, 290-292) characterizes these distinct discourse voices by their particular points of view; such viewpoints may differ, he notes, in their 'ideological' (better translated here as 'ideational') contents and in their 'axiological', or value-orientation stances toward that content and toward other voices. This is the phenomenon he terms 'heteroglossia'. I have previously tried to show how a functionalist linguistic approach to discourse analysis can make good use of these notions (e.g. Lemke 1988a, 1990, 1992). A register theory approach (Gregory 1967,

Halliday 1977) can also readily characterize differences in the typical ideational choices in two texts through the frequencies with which various options in system networks of transitivity are selected in the texts of a particular discourse voice. But to characterize their 'axiological' semantic relations, it is not sufficient to rely on the options thus far described for Halliday's interpersonal metafunction systems, important as these are. As indicated above, we do not just create relationships of offering and demanding, solidarity and distance, dominance and subordination, etc. with interlocutors; we also, and crucially for the sociologically significant meaning differences of heteroglossia, construct attitudes and evaluations toward our own and others' discourses.

In his current work on 'Appraisal,' Martin (in press) is developing system network descriptions of our semantic options for evaluating people, things, and phenomena. In my own earlier work on heteroglossia (Lemke 1992, 1995b), I had begun to identify some of the typical sorts of evaluations and attitudes we make towards facts, propositions, proposals, and discourses. After this work was completed, a corpus-analysis study by Francis (1995) came to my attention and I was struck by the almost total identity between the semantic classes her data revealed for a particular lexicogrammatical frame (see below) and those I had found in the text semantics of evaluation. I realized that what we were both seeing reflected an apparently rather fundamental principle of the semantics of English, and one that could provide the key to analyzing attitudinal meaning, at least for the evaluations of propositions and proposals so important in heteroglossia.

3.0 "It is (very) ... that ..."

Francis' (1995) data showed that if we consider occurences of sentences or clauses of the form:

(1) It is ... that ...

where that introduces an embedded (rank-shifted) noun clause, and the extraposed it is is followed by an adjective, then the adjectives which occur in this frame fall into a small number of semantic classes, all of which are in some basic sense evaluative epithets. If we apply a systemic-functional analysis to this structure (cf. Halliday 1985), we realize that it most usually occurs when the adjective is an Attribute, the noun clause its Carrier, and the noun clause represents a Fact or proposition, if realis (see Halliday 1985: 243-248), or some sort of proposal or possibility (if irrealis). So, what I believe the empirical data are telling us is that, in English at least, the only semantic attributes that propositions and proposals can have are evaluative ones. Further support for this comes from the fact that all these attributes are gradable and can be further qualified quasi-quantitatively by degrees. Francis' frame now becomes a useful test for the hypothesis and a useful heuristic for exploring the semantics of evaluation, at least in the domain of propositions and proposals:

(2) It is (Degree) [Attribute: evaluative] that [Proposition/Proposal]

e.g. It is very important that John is coming / that John come.

In the case of irrealis proposals, as we will see, English sometimes requires whether in place of that and also allows non-finite complement clauses introduced by to or for ... to.

So, what are the possible evaluative attributes of propositions and proposals?

4.0 The Evaluative Dimensions

It is useful to think of language as giving us resources for positioning propositions and proposals somewhere in a multi-dimensional semantic space. On each dimension the proposition (hereafter I will distinguish proposals only as needed) has a position that can be either positive to some specified degree, or negative to some degree. It is also a characteristic of the semantics of evaluative attributes of propositions that they are bi-polar: for every positive attribute there is a complementary negative one. I present here, with some made-up examples to simplify giving a sense of the possibilities, my best current glosses for the basic evaluative semantic dimensions (Table 1).

Table 1. EVALUATIVE ORIENTATIONS: SEMANTIC DIMENSIONS

Evaluative Attributes of Propositions and Proposals

DESIRABILITY / INCLINATION

It is simply wonderful that John is coming / that John may come.

It is really horrible that John is coming / that John may come.

WARRANTABILITY / PROBABILITY

It is quite possible that John is coming / that John may come.

It is very doubtful that/whether John is coming.

NORMATIVITY / APPROPRIATENESS

It is quite necessary that John come / that John is coming.

It is entirely appropriate that John come / that John is coming.

USUALITY / EXPECTABILITY

It is quite normal that John is coming / may come.

It is highly surprising that John is coming / may come.

IMPORTANCE / SIGNIFICANCE

It is very important that John is coming / may come.

It is really quite trivial that John is coming / may come.

COMPREHENSIBILITY / OBVIOUSNESS

It is perfectly understandable that John is coming / that John may come.

It is quite mysterious that John is coming / why John is coming.

HUMOROUSNESS / SERIOUSNESS

It is just hilarious that John is coming! / that John may come.

It is ironic that John is coming / may come.

It is very serious that John is coming / may come.

Examination of Table 1 will show that each of the dimensions could be elaborated into a large set of semantically related evaluative epithets, as well as that there are subtleties which I am not going to explore here, especially regarding irrealis proposals. Note also that, for example, a lexical item such as important, when evaluating a proposal (irrealis), such as It is important that John come, realizes a different semantic dimension from the same item evaluating a proposition assumed to be true: It is important that John is coming. Only the latter case corresponds to the semantic dimension of Importance/Significance (we do not say, *It is very significant that John come.). A single lexical item may also realize the conflation of two or more evaluations, as miraculous realizes both Unusual and Desirable, but alarming Unusual and Undesirable. A partial test for whether two items x and y share a common semantic dimension is whether it is sensible ordinarily to say: It is very (x ) but not at all (y) that .... Where this makes sense, the attributes are relatively independent, and otherwise belong to the same semantic class.

Inspection of Table 1 will also show many readers the close kinship between the semantics of propositional evaluations and Halliday's analysis of modality in the clause (Halliday 1985: 334-340; see also Martin 1995). Lexicogrammatical Modalizations for probability and usuality in the clause are one way of realizing the semantic evaluations I call Warrantability and Usuality. Modulations for obligation and inclination realize Normativity and Desirability, respectively. There are many lexicogrammatical ways to realize these semantic options, e.g.:

John must be coming. (Modal auxiliary)

John is certainly coming. (Modal adverb)

It is certain that John is coming. (Evaluative epithet/Objective orientation)

I am certain that John is coming. (Evaluative epithet/Subjective orientation)

It is a certainty that John is coming. (Evaluative nominalization)

I know that John is coming. (Modal projection: mental process)

These variants are not perfectly synonymous in their total meaning potentials, of course, but they occupy the same relative positions in systems of semantic contrast for evaluative meaning (e.g. must be/may be, certainly/probably, certain/probable, certain/of the opinion, a certainty/a good possibility, know/think). Extended discourse gives us many more means (see below).

Halliday (1985: 333) early on recognized evaluative attributes of propositions and proposals as one of the 'interpersonal grammatical metaphors' by which modality might alternatively be realized. But if we start from a strategy of identifying such attributes directly, then we find a somewhat larger set of semantic options. It is not clear that Importance, Seriousness, and Comprehensibility can have realizations as direct modifications of Process in the clause, in the same way that Warrantability, Usuality, Normativity, and Desirability can. We also find that Desirability is rather larger as a semantic class than what might be imagined from the case of Modulation:Inclination, comprising not simply what we might be keen to do, but all that we may be keen on, once done or happened. Perhaps when we are in a position to overview the evaluative semantics of actions and events, as well as of propositions and proposals, we will better understand how the basic evaluative dimensions map onto various domains of evaluands somewhat differently. It is certainly also true that the three extra dimensions are much less frequently salient in the registers I have examined compared to those which have Modality realizations.

5.0 Evaluations in Newspaper Editorials: The Corpus

Seeking a small text corpus where I would be likely to find a high density of evaluations, I chose to look at newspaper editorials. I wanted to see how comprehensively the seven dimensions could classify instances of evaluations of propositions and proposals, how frequently each dimension was realized in this typically evaluative register, what were the most typical lexicogrammatical means for expressing these evaluations, and most of all what happens when we look at evaluations in connected text.

From a larger corpus of newspaper editorials culled from what was available at the time in digital files on the Internet, I selected for close analysis 7 editorials from three newspapers: the New York Times, the Boston Globe, and the Irish Times of Dublin. These all address middle to upper-middle class readerships, and we would expect to find a wide range of lexicogrammatical and discourse devices employed, including all the various 'grammatical metaphors' for modality. This is certainly the case. Table 2 describes the relevant features of this mini-corpus.

Table 2. The Newpaper Editorials Corpus

7 editorials from 3 newspapers (NY Times, Boston Globe, Irish Times)

2676 words, 116 sentences, 218 finite clauses

418 Evaluators: 3.6 per sentence; 1.9 per finite clause

199 Desirability, Inclination 47%

71 Warrantability, Probability 17%

61 Normativity, Obligation 14%

85 Usuality, Expectedness 20%

27 Importance, Significance 6%

6 Comprehensibility, Mystery 1.4%

0 Seriousness, Humor 0

In Table 2 'evaluators' refers to lexicogrammatically segmentable elements of a text which function, as a whole, to evaluate one or more other elements along one or more of the seven evaluative semantic dimensions. An evaluator was counted only once in the totals, even if it evaluated another element on more than one dimension. There is some uncertainty about criteria for segmenting evaluators in connected text for reasons that will become clear later (see on propagation of evaluations below), but with relatively consistent criteria, the percentages for each dimension represent fairly robust results. Desirability is more than twice as common as Warrantability, Normativity, and Usuality, which occur with roughly equal frequency, and Importance occurs about half as often as they do. Comprehensibility was evaluated only sporadically in these texts, and Seriousness not at all.

The relative frequencies of some of the most common lexicogrammatical forms which these evaluators took can be seen in Table 3.

Table 3. Frequency of Types of Grammatical Realizations:

65 Attributive adjectives of quality 16%

18 Attributives from Processes 4%

91 Finite verbs (without modals) 22%

30 Modals (including emphatic polars) 7%

45 Nominalized Processes 11%

50 Abstract nouns (not from Process) 12%

10 Concrete nouns 2%

54 Complete nominal groups or prepositional.phrases 13%

37 Adverbs 9%

20 Idiomatic expresions 4.7%

Some of the categories used in Table 3 are also ones for which universally applicable criteria cannot be easily given, and some few items could have been equally well placed in another of these categories. The figures given are merely meant to be illustrative and not definitive. Most significantly, they show, I think, that there was a considerable diversity of lexicogrammatical means for evaluations employed in these texts.

There was also another, rather remarkable feature of this register. As can be seen in Table 2, there were on average about 2 explicit evaluators per finite clause. This shows the very high density of evaluations in this text. What is more remarkable is that, across all three newspapers, the range was only from 1.7 to 2.4 per clause, and within the two papers for which there are three editorials each, the figures were: 2.15, 2.3, 2.0 and 1.7, 1.6, 1.75. I think these figures shows how consistently writers unconsciously adopt the features of a register or local style, and they perhaps also indirectly argue for the consistency with which the criteria were applied in the analysis. These figures appeared only at the end, when a table was created from the raw data produced by the close analysis.

6.0 Examples of Editorial Evaluations

I would now like to offer some real textual examples of evaluations on each of the six semantic dimensions represented in the corpus. What appears when we go to actual text is that there is a tremendous interdependence of different evaluations. As Martin (1992: 553-559) and others have noted, the realizations of interpersonal meanings, including modalities and attitudes, tend to be more 'prosodic' than the more segmentable and localized realizations of ideational meanings. We can interpret this as saying that redundant, qualifying, and amplifying or restricting components of the overall evaluation are spread out through the clause, clause-complex, or even longer stretches of text. As this happens, they overlap with other evaluative meanings, and practiced writers find ways to smoothly integrate the results through delicate lexical choices and grammatical interdependencies. It will also become clear that evaluations of propositions and proposals are not independent in connected text from evaluations of the participants, processes, and circumstances within propositions or proposals. The evaluators counted in Tables 2 and 3 include those which evaluate most directly a semantic element of a proposition (other than themselves) as well as a whole proposition. Since nominalizations of various sorts are also common in this register, what is a proposition at one point in a text readily becomes 'condensed' (see Lemke 1990) as a participant at another, and participants (especially abstract nominals) are often meant to be correspondingly 'expanded' by the reader into implied propositions through reference to some known intertext, as well as through reference to the immediate co-text.

6.1 Desirability:

(3) FORTUNATELY, the Senate has shown that it can rein in THE EXCESSES OF the House. (Contract 4.1)

Evaluators are shown in upper-case. (Citations to the corpus are by nickname of the editorial, followed by consecutive paragraph number, and consecutive sentence number within paragraph. See Appendix for sources of the editorials.) Here we have a sentence adverb evaluating on Desirability the entire proposition which follows. In context,

excesses also carries forward the evaluation of negative Desirability; it is an evaluative nominal, conflating negative Desirability with its ideational meaning. It can also be expanded, from co-text and known intertexts, into a proposition about what the House has done. Note also that it can serve as an evaluative epithet directly in our test frame: It is really quite excessive that .... I have not highlighted here shown, which is an evaluator on another dimension (Warrantability).

6.2 Warrantability:

(4) Western experts BELIEVE that the largely untapped oil and natural gas riches of the Caspian Sea countries COULD MAKE that region the Persian Gulf of the next century. (Game 1.2)

Here we have a mental process believe projecting a proposition and at the same time evaluating it for Warrantability. The same evaluation is then carried forward within the proposition by the epistemic modal auxiliary could within the verbal group. There are also evaluations of Desirability here, by lexical choice (untapped, riches) and by intertextual connotation (Persian Gulf).

6.3 Normativity:

(5) At the same time, change is urgently NEEDED to ALLOW it to fill a dangerous vacuum. (North 1.4)

This sentence presents a double evaluation of Normativity: what is needed is that something (recoverable by cohesive reference) change, and what is allowed is that something (again recoverable) fill the vacuum. We find a characteristic Degree of enhancement for the first evaluation in urgently. There is also clearly an evaluation of Desirability by dangerous (cf. It is very dangerous that this vacuum exists.)

6.4 Usuality:

(6) SURPRISINGLY, despite all the caterwauling from ..., the party's MAINSTREAM MAJORITY may slowly be bending G.O.P. policy to fit their libertarian views on abortion. (Dole 1.2)

The sentence adverb again evaluates, this time for Usuality (negatively, i.e. contrary to what is usual or expected), all that follows. Within the proposition, mainstream majority resumes the theme of what is or is not usual, but in a separate evaluation, ostensibly conflated with the Actor within the main finite clause, yet extending semantically to the non-finite proposition (via the cohesive link their), semantically akin to: In the majority of cases (i.e. usually) Republicans (members of the G.O.P.) hold libertarian views on abortion.

6.5 Significance:

(7) To say this is not to deny THE SIGNIFICANCE OF symbols; only to question the extent to which they should BE CENTRAL TO the debate about effective policing in the light of negotiated change in political and social structures. (North 5.2)

Here we again find the attribute nominalized and transferred from a proposition (that certain things are important as symbols) to one participant symbols. This theme is then picked up in a later clause in which the idiom be central to functions as a locative metaphor for the evaluative attribute Important.

6.6 Comprehensibility:

(8) There are NO EASY ANSWERS to the problem of policing arrangements in the North. (North 1.1)

The examples of Comprehensibility in this small corpus (there are only six candidates) are all rather indirect ones. Reference is made to 'shadowy arrangements' to things being 'complicated' and here to a lack of 'easy answers'. My interpretation in these cases is that part of what is being meant is that a situation or proposition is not obvious and easily comprehensible but in some way unclear or mysterious. In our test frame, we can certainly expect forms like It is obvious that ... and I am fairly confident of including this dimension. Francis (1995: 53) also finds examples and alots them their own gloss ('obviousness'). In this particular example we see the pervasive tendency for metaphorical transfer among the evaluative semantic dimensions and, here, the closely related semantics of Ability. What is being evaluated is the problem, which here is again a nominalization expandable as a proposition (and already including a negative evaluation of Desirability). The answers or solutions are not easy (sc. to find) when the problem is difficult or complicated, the situation problematic or unclear. The answers are not obvious because the situation described by the (implied) proposition is not.

6.7 At the borders: Ability, Seriousness, Temporality and Temporariness

I will not make any further comments about the rarest dimension, seriousness/ humorousness. It seems to be something of a marginal or borderline case, with relatively few lexical realizations that fit the test frame and some awkwardness with many following propositions. It did not occur in the editorials corpus. There is also an interesting borderline case, which I take to be, like Ability, not an evaluative attribute of propositions as such, but which seems to be intimately related in its semantics.

Ability is listed by Francis (1995) on purely statistical grounds as one of the classes of adjectives in the frame It is (possible, difficult, easy) to .... It does not occur in the same semantic sense when followed by Fact clauses (i.e. is to be distinguished in meaning from It is possible that ...). What is formally similar is here semantically quite different in that Ability is an evaluative attribute of Persons or Things as agents and not of propositions or proposals as such. It is possible for John to come or It is possible to see John coming are evaluations of the Ability of John or of the speaker to do something, and not evaluations (as in It is possible that John is coming) by the speaker of a proposal or proposition about John. It can contribute only indirectly to such evaluations (see below).

The more nearly borderline case which I have identified from analysis of the corpus occurs 14 times, distributed among 5 of the 7 editorials. It often collocates with Usuality judgments, and I propose to call it Temporality or Temporariness. These are judgments that states of affair have changed, are newly arisen, or only occur momentarily -- and so

cannot be taken as Usual. In some cases these locutions clearly evaluate something as Not-Usual. In other cases they only imply this rather indirectly. For example:

(9) Proudly, justly, Palestinians have long insisted that ONCE THEY WERE FREE of Israeli occupation they intended to construct a polity different from the police states surrounding them in the Arab world. (PLO 1.1)

Here have long insisted does evaluate for Usuality, at least in an aspectual sense of habitual action, and in once they were free we sense, here still as irrealis, a punctual temporality, a possible change or new beginning, something that would clearly not correspond to the Usual state of affairs in the judgment of the writer. Note that the theme of what is or would not be Usual in the Arab world is taken up again with different from ... surrounding ....

(10) The Republicans' cramped vision of limited government has won, FOR NOW. (Contract 1.7)

In this example, for now is focal and emphatic, and clearly implies a temporariness for the state of affairs being evaluated, as if to say: It is only temporarily that .... Not a precise fit with our test frame, but rather close. I think we need a fuller understanding of how the semantics of Usuality relate to those of Aspect and Time generally.

7.0 Evaluations and Text Semantics

As I have argued elsewhere (Lemke 1988a, 1988b, 1995b) when we consider the meanings made by extended, cohesive texts that are not made in individual clauses, we often encounter phenomena of language that reveal new semantic resources at the text level. This is particularly true for the semantics of evaluation because of its tendency toward 'prosodic' realizations, i.e. realizations that tend to be distributed through the clause and across clause and sentence boundaries. We have already seen some limited examples above. I would like to note two kinds of text semantic phenomena in the editorials corpus: evaluative metaphor and the textual propagation of evaluations.

7.1 Evaluative Metaphor and Prosodic Overlap

When I say It is important that John come, I am not mainly speaking about the Significance of an act, but trying to convey its Normativity: Necessity (in a milder degree). It is important that John come is far closer in meaning to John really must come or We need John to come than it is to The fact that John is coming is quite significant. The lexical item and its ideational meaning are only the gateway here; having entered it, we are swept along a metaphorical transference from the expected dimension of Significance to the actual evaluation of Normativity. There are many other cases in which we fairly liberally substitute one kind of evaluation for another, when the context makes clear which one we are really getting at. But this simple metaphoric process (one it seems to me rather on the border between lexical metaphor and grammatical metaphor; cf. Halliday 1985: 320) becomes in practice quite complicated as we examine how prosodic overlap between different evaluative themes in a text facilitates the shifting and overlaying of different evaluative meanings. I will offer two examples.

(11) But in an era when candidates are too often marketed like toothpaste, campaign discourse can descend to the level of hucksterism. (Reckless 2.1)

How are we to interpret the evaluative force of too often here? The basic lexis invokes the semantics of Usuality: it is usual, the writer says, that candidates are marketed like toothpaste. The Degree, too, however, carries an ideational sense of Excess, which is implicitly evaluative for negative Desirability. More fundamentally, we observe in our culture a basic logical principle that if something is bad, more of it is worse. That marketing candidates like toothpaste is bad is construed by the extension back to this clausal proposition of the very strong negative Desirability evaluators in the following clause: descend, hucksterism. The ideational link between the clauses is not very explicit; one needs some interpretive skill to identify campaign discourse with marketing candidates (the missing links are supplied by way of collocations specific to relevant intertextual thematic formations, cf. Lemke 1983, 1985). But in terms of evaluational meanings, there is first the contrast in Importance that can be assumed between candidates and toothpaste, and while the latter may not be intrinsically un-Desirable, identifying something Important with something so un-Important potentially is, and this potentiality is activated in the text by the overlapping un-Desirability of an excess, signalled by too. Prosodically we hear a sustained tone of un-Desirability which rides at one point on top of a high degree of Usuality. Each reinforces the other. As we listen more closely we realize that in an era already begins to invoke Usuality, that what is Usual here is also un-Desirable, and the fact that it is very Usual makes it even more un-Desirable. The salience of the ideational meaning of Usuality, that something is simply Frequent, is rather low here compared to the dominant tone of un-Desirability. We could say, and similar cases abound in the corpus, that Usuality here is functioning as a kind of close grammatical metaphor for Un-Desirability, or perhaps more precisely that insofar as the two kinds of evaluative meanings overlap and support one another textually, while the meaning of Usuality remains, its evaluative force contributes also to producing un-Desirability of a high degree.

(12) A compact that addressed those concerns, while equitably sharing profits among the Caspian states and foreign investors, could make all a winner in this complicated game. (Game 6.1)

There are a lot of evaluations going on in this single sentence (Desirability, Warrantability, Comprehensibility, perhaps Normativity). If we focus on the modal could and ask about its evaluative force, what it tells us about how the writer seems to view the warrantability, desirability, etc. of the matters written of, we see first of all a middle degree of Warrant: this proposal could be true, is probably true (the degree, or Value, in Halliday's terminology for modalization would be Low on purely lexical grounds, but is raised by contextual factors I do not discuss here). But a competent reader of this text is also going to understand that what is going on here semantically is not primarily an estimate of probability but a proposal for desirable action. The complete verbal phrase here is an idiom: could make all a winner, and winner marks the Desirability and all amplifies its degree. So we have also a second interpretation, that could stands here in its ideational role of realizing Ability, sc. that of the compact being proposed. The prosody of Desirability here rides on top of the semantics of both Ability and Warrantability. If the compact is able to do it, that's desirable. If it's probable that the compact will do it, that's Desirable. And the compact and the implied proposition about it becomes more Desirable insofar as its Ability to perform as claimed is the more

highly warranted as true or likely. One could say here that Ability and Warrantability are functioning as metaphors for Desirability of the implied action, but as we see the full situation is more complicated.

7.2 The Phenomena of Evaluative Propagation

We have already begun to see examples of what is to me the most fascinating of the text semantic phenomena surrounding evaluative meanings. If we consider evaluators as semantic operators, and ask what is their scope, i.e. to what precisely does their evaluation extend -- we find that quite often evaluations propagate or ramify through a text, following the grammatical and logical links that organize it as structured and cohesive text as opposed to a mere sequence of unrelated words and clauses.

As Hasan (1984, 1989) has emphasized, connected text is organized by two complementary principles, which she has termed 'structure' and 'texture'. In the former case we find multivariate patterns where the juxtaposition of functionally (and sometimes also formally) differentiated parts A--B--C--D comprise a whole. The functional and syntactic structures of the clause or nominal group are common examples. In the case of texture, a number of elements, otherwise isolated, are linked through a text to form a 'chain' by virtue of their semantic relations to one another (such as co-reference, synonymy, etc.). As we will see, evaluations can propagate from one element of a structure to another, as well as along cohesive chains, but they can also create cohesive chains.

7.21 Evaluative Cohesion

Co-evaluation, along the same dimension (and more definitively if also similar in polarity and degree, but this is not necessary), can create cohesive links between separated elements of a text that are not readily construed by the usual cohesive devices. For example:

(13) Meanwhile, A GROWING NUMBER of Republican politicians, including Christine Todd Whitman and William Weld, have developed the stiffer spines that were OUT OF FASHION among Eastern Republicans during the Bush years. (Dole 4.3)

I believe that part of the cohesiveness of this text is contributed by the fact that both a growing number and out of fashion construe meaning on the evaluative dimension of Usuality. In fact, they are in semantic contrast in this text, even though they are not lexically in any sense antonyms. The first contributes to an evaluation of Usuality:Frequency, and the second to Usuality:Expectability.

7.22 Syntactic Propagation

Within the clause evaluative stances toward one structural element (participant, process, circumstantial) may transfer their evaluation to another element. If we exclude the explicit evaluators which work this way (attitudinal Attributes/Epithets which evaluate their Carriers/Things, auxiliaries which modalize their Head verb, and process nouns derived from verbs of explicit evaluation), there are still a host of other phenomena. It is particularly interesting that the Polarity of these evaluations can be reversed during

propagation (Degree normally propagates unchanged) by a large number of very common locutions. Consider, for example:

(14) Since 1980, the Republican platform has opposed abortion without exceptions and promoted an ammendment to ban the procedure. (Dole 3.2)

The evaluations within this text depend on a single variable which must be assigned intertextually: the Desirability of abortion. The rest of the elements of the text received evaluations on the dimension of Desirability indirectly once a value is assigned to this keystone. If we take abortion (or here the right to it) to be Desirable, then to oppose something Desirable is un-Desirable (propagation from Range to Process, with reversal of polarity), and the more so with a high Degree (without exceptions). But the propagation does not stop with the Process, for whatever opposes the Desirable is itself to that extent un-Desirable, i.e. the Republican platform. If we start from the co-referent the procedure, we have a propagation by cohesive tie from abortion, and so must consistently take the procedure to be Desirable, hence banning something Desirable would be un-Desirable (Range to Process with reversal), what does this banning is also un-Desirable (Process to Do-er), sc. an ammendment. Promoting this undesirable thing is un-Desirable (Range to Process, no reversal), and so finally once again, what promotes is un-Desirable, namely the same Republican platform.

It is quite common that a reader needs intertextual knowledge of the writer's probable assignment of value polarity to key well-known elements in order to trace out the evaluations in the text. In the case above, one logically assumes that the Republican platform evaluates abortion as un-Desirable. Inter- and co-textually it is clear that the editorialist evaluates it as Desirable, and so these counter-evaluations are also functioning to set up a 'heteroglossic opposition' (Lemke 1995b) between the writing and referenced discourse voices. It is also quite common that it is the ideational semantics of processes which determines whether they can propagate an evaluation and whether the polarity will reverse or not. So to increase, amplify, support, promote something Desirable, is Desirable, and to decrease, oppose, resist, limit it is un-Desirable, etc.

7.23 Projective Evaluation

(15) Thomas Parker, a lecturer in education at Boston University, predicts a dramatic rise in self-learning and ... (Education 5.1)

In this very typical example, what is at stake is Warrantability, the writer's view of the probability of something someone else is saying. Lexically predicts is a projecting process (verbum dicendi, cf. Halliday 1985: 196-198) that has relatively low warrantability; we assume that predicting the future is at best an uncertain matter. But the reliability of predictions, and so the Warrantability of the proposition predicted, rises if we specify a credible Sayer for the process, here a lecturer in education at Boston University. Note that the force of credibility arises intertextually and heteroglossically (what is the default credibility of university lecturers, of lecturers in education, of lecturers at Boston University?) but also internally and ideationally insofar as a lecturer (cf. expert) in education is Sayer for a prediction about self-learning (as opposed to, say, about soybean futures). Here and elsewhere the evaluation of the Sayer as reliable (perhaps the analogue for evaluations of Persons to Warrantability for propositions)

propagates to increase the Warrantability of the proposition projected, and characterizing the Sayer as unreliable decreases it. The ideational meaning of the projecting verb, determining a default Warrantability for what is projected, interacts with the reliability of the Sayer and the Warrantability of the proposition considered in isolation from the projection. Sayers associated with intrinsically Warrantable Sayings gain in reliability, credibility, and vice versa.

Projective evaluations can carry across several clauses or sentences, but we will consider here some more interesting cases of propagation of evaluations across more extended text.

7.24 Extended Prospective and Retrospective Evaluation

(16) Mr. Dole's appearance on "Meet the Press" last Sunday was NOTABLE for two things. (Dole 2.1)

Here the evaluation is on the dimension of Importance, and sentence 2.1 both sets up a cataphoric reference chain (two things) and prospectively evaluates the future contents of that chain, which extends over the following three sentences, as Important (notable).

(17) The pettiness of this attempt to bully the Palestinian press into the subservient posture of a propaganda organ for Arafat's cult of personality belies earlier hopes and promises. (PLO 6.1) Sadly, it has become part of A PATTERN OF authoritarian governance. (PLO 6.2)

As we initially read the first sentence, its evaluative concerns are tightly focussed on Desirability. We see the use of belies to reverse the polarity of hopes and promises in a syntactic-structural propagation that makes the pettiness not only undesirable in itself, but un-Desirable on this second count as well. It in turn of course functions mainly as evaluator of this attempt to bully ..., and there is no relief from the negative tonality except at the end (hopes and promises), and there only to reverse back on the core evaluand. The second sentence begins with a continuation of this negative Desirability (sadly), but suddenly we find a Usuality evaluator (a pattern of) which operates through the syntax of Identification (has become) to evaluate it as Usual, and this retrospectively ramifies up back up the reference chain once again to the primary evaluand, this attempt to bully .... Note that grammatically it ought to refer to pettiness, as belies ought also to propagate syntactically to it, but it is our sense of the evaluative meanings here, of what is being evaluated, that disambiguates the anaphoric reference. All the Desirability evaluators point to a single evaluand, and it is not pettiness, which is itself one of these evaluators. So we take the Usuality evaluation to operate on that same evaluand, even retrospectively and across a sentence boundary.

While my examples have often featured Desirability evaluations, I hope it is clear that all of these phenomena occur for all the evaluative dimensions. There are also much longer range cohesive propagations, as well as structural ones that depend on relations among elements of genre structure or rhetorical units and which can span very long distances in the text.

8.0 Conclusion

I have not attempted here to present either a fully systematic or an exhaustive account of evaluative meaning resources, even for propositions and proposals. That would be a very large subject indeed, as I hope I have suggested here. My aim has been to describe some of the interesting phenomena I have encountered and to stimulate others to carry forward the more painstaking and thorough studies required to enhance our knowledge of evaluative meaning resources.

I believe that this area of investigation is important and will be rewarding, not simply as an area of the grammar and its semantics until now relatively little explored, nor even as a revealing example of how orientational-interpersonal meanings are realized differently in text than more familiar ideational ones, but because of its significant applications in discourse analysis. We know that it is not just what people believe, but also the values we hold about what we believe which tend to shape our actions. We also know that an understanding of micro- and macro-social relations through the analysis of texts and discourse requires that we be able to map out the heteroglossic relations among the different discourse voices and valuational viewpoints in a community. Finally, it is certainly reasonable to suppose that part of what we do when we mobilize the evaluative resources of language is to help constitute our own identities, and the identities of others, as agents who believe and doubt, desire and detest, and judge importance, appropriateness, usuality, comprehensibility, and seriousness. For all these reasons we need a more adequate linguistic account of this critical function of language: evaluation.

Appendix: Sources of the Editorials

The New York Times

Contract = "The Contract, Partly Fulfilled", August 20, 1995

Dole = "Call Him Mr. Flexible", December 20, 1995

Game = "The New Great Game in Asia", January 2, 1996

The Boston Globe

Education = "Upgrading Public Education", January 4, 1996

PLO = "The Unliberated PLO", December 28, 1995

Reckless = "Reckless Attacks on Start II", January 4, 1996

The Irish Times

North = "Policing the North", January 4, 1996.

References

Bakhtin, M. (1935) Discourse in the novel. In M. Holquist (ed.) The Dialogic Imagination (1981). Austin TX: University of Texas Press.

Francis, G. (1995) Corpus-driven grammar and its relevance to the learning of English in a cross-cultural situation. In A. Pakir (ed.) English in Education: Multicultural Perspectives. Singapore: Unipress.

Gee, J.P. (1990) Social Linguistics and Literacies. London: Falmer Press.

Gregory, M. (1967) Aspects of varieties differentiation. Journal of Linguistics 3: 177-198.

Halliday, M.A.K. (1977) Text as semantic choice in social context. In van Dijk, T. A. and Petöfi, J. (Eds.) Grammars and Descriptions. Berlin: de Gruyter.

Halliday, M.A.K. (1985) An Introduction to Functional Grammar. London: Edward Arnold.

Hasan, R. (1984) Coherence and Cohesive Harmony. In J. Flood (ed.) Understanding Reading Comprehension. Newark, DE: International Reading Association.

Hasan, R. (1989) The structure of a text. In M.A.K. Halliday and R. Hasan, Language, Context, and Text. London: Oxford University Press.

Lemke, J.L. (1983) Thematic analysis: Systems, structures, and strategies. Semiotic Inquiry 3(2): 159-187.

Lemke, J.L. (1985) Ideology, intertextuality, and the notion of register. In J.D. Benson and W.S. Greaves (eds.) Systemic Perspectives on Discourse. Norwood, NJ: Ablex.

Lemke, J.L. (1988) Discourses in conflict: Heteroglossia and text semantics." In J. D. Benson and W. S. Greaves (eds.) Systemic Functional Approaches to Discourse. Norwood, NJ: Ablex Publishing. 29-50.

Lemke, J.L. (1988b) Text structure and text semantics. In R. Veltman and E. Steiner (eds.) Pragmatics, Discourse, and Text . London: Pinter. 158-170.

Lemke, J.L. (1990) Technical discourse and technocratic ideology. In M. A. K. Halliday, John Gibbons, and Howard Nicholas (eds.) Learning, Keeping, and Using Language: Selected Papers from the 8th AILA World Congress of Applied Linguistics, Vol. II. . Amsterdam: John Benjamins. 435-460.

Lemke, J.L. (1992) Interpersonal Meaning in Discourse: Value Orientations. In M. Davies and L. Ravelli (eds.) Advances in Systemic Linguistics: Recent Theory and Practice. London: Pinter. 82-104.

Lemke, J.L. (1995a) Textual Politics: Discourse and Social Dynamics. London: Taylor & Francis.

Lemke, J.L. (1995b) Intertextuality and text semantics. In M. Gregory and P. Fries (eds.) Discourse in Society: Functional Perspectives Norwood, NJ: Ablex Publishing. 1995. [pp.85-114]

Martin, J.R. (1992) English Text. Philadelphia: John Benjamins.

Martin, J.R. (1995) Interpersonal meaning, persuasion, and public discourse. Australian Journal of Linguistics 15: 33-67.

Martin, J.R. (in press) Beyond exchange: Appraisal systems in English. In S. Hunston and G. Thompson (eds.) Evaluation in Text. London: Oxford University Press.

DISCOURSE AND SOCIAL THEORY [Sections relevant to the argument of the book, but not to our SCCS dialogue are omitted. The first section below sets the stage for the discussion of Bakhtin, Foucault, Halliday, Bernstein, and Bourdieu. Selected references are given at the end; the complete references are on my website http://www-personal.umich.edu/~jaylemke/tp-biblio.htm ]  

Social Theories of Discourse If discourse plays a critical role in social dynamics, then social theories about discourse should point the way to a dynamic, critical, unitary social theory. Unfortunately, most theories of discourse are not social theories. Indeed most theories of discourse are mainly linguistic and psychological, paying relatively little attention to the question of who says what when, why, and with what effects. The social context of discourse, and issues of discourse as social action are largely ignored. Instead discourse is mostly seen as the product of autonomous mental processes, or it is simply described as having particular linguistic features. Why is this? Granted, some people are simply interested in linguistic description for its own sake, and others want to use discourse as a tool for understanding what they call the mind. But why are our theories of linguistic description, and our theories of mind, ones that ignore the social functions of language, the social origins of human behavior, and the social position of the linguist or psychologist? The answer, I believe, lies in the ideological functions of the discourses of psychology and linguistics in our own society and its history. Social perspectives on any human phenomenon are potentially dangerous to the interests of power.  In modern times, in European cultures, we have preferred theories that claim to be universal, theories that do not admit that they may see the whole world, but can only see it from one culture's viewpoint. We have constructed a notion of "human nature" based on our own views of what is worth paying attention to in the activities of humans. We have rooted our psychology in a fanciful connection to biology and the unity of the human species. We have rooted our linguistics in this psychology. We have taken our modest successes in the atypical domains of physics and chemistry (where the objects of interest do not have the kind of complexity for which cultural differences in viewpoint can matter very much; cf. Harding 1986; Lemke 1993b; Salthe 1985, 1993) and used this to make plausible our impossible claims about the universality of our views of language and mind.

 We have not questioned the fundamental assumptions of our own cultural tradition: whether an objective science of matters human and cultural is possible in principle, whether the notion of mind as associated with both a biological organism and a social person is tenable, whether social systems can usefully be thought of as being composed of individuals as such, whether our subjective experience of ourselves as actors and perceivers is the product of the discourses and practices of our culture rather than a universal human given. Some, but not many, have asked whether our taste for universalizing theories may have arisen from the need of European societies to justify their domination of other cultures by force in the past few centuries. Or from the need of upper-middle class, middle-aged, European males to legitimate their domination by force of workers, peasants, women, children, elders, slaves and various cultural Others in their own society. It is not just common sense, but science as well, especially the sciences of the human and near-human, which we must subject to skeptical, critical examination to determine their ideological biases.  ? Our task here is to identify ? the major theories of discourse which have emphasized its social dimensions. I want to discuss particularly the work of Mikhail Bakhtin, Michel Foucault, Michael Halliday, Basil Bernstein, and (though not a discourse theorist) Pierre Bourdieu. Each of them seems to have arrived at what I see as the same basic solution to the problem of connecting discourse to Discourses, events to larger social relations and processes. Each has also contributed greatly to our resources for analyzing the social functions, including the ideological functions, of discourse. Bakhtin and Heteroglossia I begin with the work of Mikhail Bakhtin (esp. 1929, 1935, 1953) in part because he was the first of these five to try to construct a social theory of discourse, and so his work seems to us today the most original, even idiosyncratic. He worked as part of a group of scholars in the period immediately following the Russian Revolution, a time when Marxist ideas were widely respected,  and when there was a temporary crack in the monolithic ideology of European culture. In this period, Vygotsky (e.g. 1963) began to ask about the social origins of mind, standing the received wisdom of psychology on its head. Bakhtin, along with V.N. Voloshinov, P.N. Medvedev, and others wanted a theory of language and literature that saw it too as having a social origin and character, and not as being merely the autonomous product of individual minds. What, for Bakhtin, are the fundamental elements of language as a social phenomenon? Words? Sentences? Speakers? None of these: "The actual reality of language/speech is not the abstract system of linguistic forms, nor the isolated monologic utterance, nor the psychophysiological act of its implementation, but the social event of verbal interaction implemented in an utterance or utterances." (Bakhtin-Voloshinov 1929/1986: 94) An utterance, a moment of discourse, as a social event, as an act that contributes to the social activity of discourse: this for Bakhtin is the starting point. But what of the meaning of this event? For Bakhtin our meanings do not arise in individual acts of will

in which we are the sole determiners of our utterances, because a verbal act "inevitably orients itself with respect to previous performances in the same sphere, both those by the same author and those by other authors" (p.95). The utterance always originates in and functions as part of a social dialogue (whether the other participants in this dialogue are considered to be actually present or are only implied): "The linguistic significance of a given utterance is understood against the background of language, while its actual meaning is understood against the background of other concrete utterances on the same theme, a background made up of contradictory opinions, points of view, and value judgments." (Bakhtin 1935/1981: 281). This is a view of meaning that came later to be called the principle of intertextuality (cf. Kristeva 1980; Lemke 1985, 1993d), because it sees the meaning of each particular utterance or stretch of discourse as arising in the relations between sayings and social viewpoints, and not in relations among linguistic forms as such, or among speakers as individuals. We make sense of every word, utterance, or act against the background of (some) other words, utterances, acts of a similar kind. This implies, of course, that it is very important to understand just which other texts a particular community considers relevant to the interpretation of any given text. In what he says, Bakhtin distinguishes between a narrower, formal linguistic, or semantic view of meaning and a broader more social view. The former depends on features of the language itself, and we will later call it the semantic meaning potential of the utterance as a linguistic form. It tells us what this utterance could mean, across a variety of contexts, insofar as it is interpreted consistently with very general principles of grammar and word meaning. But the latter is what the utterance actually does mean, as a social act, in the context in which it is used here and now. And that in turn depends on a whole social system of utterances made in various times and places, a system of texts written or said from different viewpoints, embodying different opinions and values. The notion of the utterance for Bakhtin is a bridge between the linguistic and social, the event-meaning and the larger social systems in which that event has its meaning for us. Bakhtin went on to develop this view of the utterance into a more general view of discourse as always implicitly dialogical , as always speaking against the background of what others have said or written in other times and places. He describes the struggle to make a word or utterance one's own, to place it in a new context as a new social event, so that its meanings are as much our own as another's. Along the way he began to see that the background against which an utterance means is not simply a set of isolated, unrelated utterances. He saw the diversity of language, how the utterances of people from different times and places, and different social positions, were systematically different: "Language is unitary only as an abstract grammatical system of normative forms, taken in isolation from the concrete ideological conceptualizations that fill it. ... Actual social life and historical becoming create within [a language] a multitude of bounded verbal-ideological belief systems ... [within which] are elements of language filled with various semantic and axiological content, and each with its own different sound." (p.288) 

What he calls at this point "bounded verbal-ideological belief systems" he elsewhere glosses as the "social languages of heteroglossia" or as distinct social voices . He illustrates what he means by referring to the stratification of language in actual use into a variety of "social class dialects, languages of special groups, professional jargons (including those of lawyers, doctors, teachers, and novelists), genre languages, the languages of generations and age groups, of the authorities, of literary and political movements, historical epochs, etc." (1935/1981: 262-3, cf. 289). "All the languages of heteroglossia ... are specific points of view on the world, forms for conceptualizing the world in words, specific worldviews, each characterized by its own objects, meanings, and values. As such they may all be juxtaposed to one another, mutually supplement one another, contradict one another, and be interrelated dialogically." (1935/1981: 291-2). Here Bakhtin is articulating his critical insight that the various social voices, the various characteristic discourses of different social groups, have specific, ultimately sociological, relations to one another. All the social relations of groups, their alliances of mutual support, their conflict in opposition to one another, are created, re-created, negotiated, and changed in the social dialogues of our discourse with one another. What Bakhtin calls social languages or voices we have been calling Discourses, or now more formally, discourse formations . These are the persistent habits of speaking and acting, characteristic of some social group, through which it constructs its worldview: its beliefs, opinions, and values. It is through discourse formations that we construct the very objects of our reality, from electrons to persons, from words to "discourse formations". And we necessarily do so from some social point of view, with some cultural system of beliefs and assumptions, and some system of values, interests, and biases. But we do this not as individuals alone, but as members of communities, and however we do it, whatever discourse formations we deploy to make sense of the world, our formations always have systematic sociological relations to their formations. We speak with the voices of our communities, and to the extent that we have individual voices, we fashion these out of the social voices already available to us, appropriating the words of others to speak a word of our own. In the theory of heteroglossia, all the key elements of a social theory of discourse are present, including a dynamic model: "Language and languages [i.e. heteroglossic discourse types] change historically primarily by means of hybridization ... the crucible for this mixing always remaining the utterance." (1935/1981: 358-59). The notion of hybridization is that particular utterances, even though the product of a single speaker, may contain within them elements of more than one dialect or discourse formation, thus producing new possibilities, which, if taken up by other speakers, can lead to linguistic and cultural change. How has Bakhtin built his bridge between the event (the utterance) and the social system of heteroglossia (the social relations of various constituent groups in a society)? First by the principle of intertextuality: that the meaning of an utterance or event must be read against the background of other utterances and events occurring in the

community. And second by introducing an intermediate notion between the social event and the system of social relations: the social language or voice characteristic of a particular group in the community. The principle of intertextuality needs to be further specified. We need to understand just how members of a community read one text against the background of some, and not other, texts to construct its meaning. And the principle of heteroglossia will need ultimately to tell us more both about how different social groups come to speak and act differently, and about the relations between the discourse habits of a group as such and the discourse habits associated with the various activities in which members of the group engage. But Bakhtin's principles are foundations on which we can build such a social theory of discourse. I doubt that I would have recognized the significance of these principles when I first read Bakhtin in the early 1980s if it had not been for the familiar ring they had. I had already encountered, I realized, these same principles, in different terminology, in the work of Halliday, Bernstein, Foucault, and Bourdieu.  Halliday and Bernstein: Register and Code  Bakhtin's notion of the social languages of heteroglossia was modeled on the diversity of the regional and social dialects of Russia in his time, and so for him these forms of discourse were associated specifically with the groups of people who used them. The British linguist Michael Halliday, some forty years later, was trying to describe the linguistic differences associated, not with different communities of speakers, but with different activities in social life. We all recognize, as did Bakhtin, that the language of mathematics is different from the language of sports or politics. Halliday sought to characterize these differences more specifically, or, as he would say, more delicately. Unlike Bakhtin, however, Halliday had at his command a very powerful semantic analysis of the grammar of his own language (English). He recognized that the language of a sports report, a sales transaction, and a newspaper editorial differed not simply in their vocabulary, and not simply because these uses of language are more likely for people in some social positions than others, but because the frequencies of occurence of many grammatical and semantic features in these texts were skewed by the nature of the different activities in which language was being used. From this came his now well-known theory of registers : the functional varieties of language, characteristic of particular activities in which language is used, defined by systematic differences in the probabilities of various grammatical and semantic features in the texts of each register (Halliday 1977, 1978; see also Gregory 1967). Where the field of the activity differed, as say between politics, sports, or mathematics, there were characteristic differences in the frequencies of say action verbs vs. relation verbs, or active vs. passive voice; where the tenor of interpersonal relationships (including intimacy and power relations) differed, there were corresponding differences in mood (interrogative requests vs imperative commands, say) or in modality (simple polar verbs vs. modal auxiliaries indicating possibility or doubt); and where the differences were those of mode , as between speech and writing, or the language of participation vs. that of observation, there were differences in how information in one clause was highlighted

or backgrounded and linked to information in other clauses (thematization, cohesion, etc.). Though register theory was initially only about differences, about variation in linguistic features from one sort of activity or situation type to another, people quickly found it useful to speak of the register of this or that activity. More delicate analysis (e.g. Gregory & Malcolm 1981) showed that while Halliday's arguments apply statistically to the whole of a text, that within a text, as we move from one section to another, there is smaller-scale (phasal ) variation in how the text constructs its meanings. Texts have internal semantic structure, which further reflects the detailed functions of each particular stage in the activity that gave rise to the text, or which the text is describing or enacting. Halliday, along with Ruqaiya Hasan (e.g. Halliday & Hasan 1989, Hasan 1984b) and Jim Martin (1985, 1992), have since tried to work out more detailed connections between register variation and the internal structures of texts of different kinds. These kinds, or genres , also identified by Bakhtin (1953), from familiar literary ones such as sonnets and folktales (e.g. Propp 1928) to expository genres like the scientific research article (e.g. Bazerman 1988) to spoken genres such as those characteristic of the dialogue of the sales transaction (e.g. Mitchell 1975, Ventola 1987) or the dialogue of the classroom (e.g. Lemke 1990) are again all characteristic of activities rather than of groups of people as such. Halliday's social theory of discourse suggests that our uses of language are inseparable from the social functions, the social contexts of actions and relationships in which language plays its part. Halliday suggests that language be viewed as a system of resources, a set of possible kinds of meanings that can be made, and that we then examine which kinds of meanings actually get made in the course of which human activities, by which social participants. This is what is meant by seeing language as a social semiotic , a resource to be deployed for social purposes. This view is quite consistent with the key principles we have identified from Bakhtin. It makes it possible to identify a number of the grounds on which a community may find one utterance or text relevant for the meaning of another (that it is of the same register, or the same genre; that it was constructed in the course of the same kind of activity, etc.; cf. Lemke 1985). It also introduces an intermediate notion between the text or utterance, and the social system: the system of registers and genres in a community. Implicitly it shifts the emphasis toward seeing the fundamental elements that define the community as its system of activities or social practices, rather than viewing it directly as a system of different types of individuals. But there is obviously one link missing: how are we to understand the differences in language-using habits between those of different ages, genders, social classes, subcultures, etc.? Halliday was greatly concerned with this question, and in the 1960s and 1970s both he and Ruqaiya Hasan collaborated with Basil Bernstein, a sociologist working in the field of education, in order to forge this missing link. Bernstein (1971, 1975) called it code , or later, semantic coding orientation. It was greatly misunderstood in its day, especially in the United States, where great efforts were being made at that time to show that all social dialects, especially those of oppressed African-Americans, were powerful resources for meaning-making, and not merely clusters of random

mistakes in grammar. Bernstein tried to point out something that is now largely taken for granted: that the schools expect people to use language in certain ways, and that these are by and large the ways of the upper-middle class, putting the members of other social classes at an automatic relative disadvantage. Bernstein argued, as has now been quite well established by the later work of Hasan (e.g. 1989b, Hasan & Cloran 1990), that the communities formed by members of different social classes learn to use language differently, so that even in what seems to be the same social activity (say, mothers questioning or scolding their children), even after we have taken register difference into account, there are further differences in the frequencies and characteristic combinations of grammatical and semantic options that are taken up by members of different social classes. Hasan has shown similar sorts of difference according to gender as well (e.g. 19..). These are not small differences. They stand out in plots of Hasan's data so strikingly that statistical tests of their significance are hardly necessary (though of course they have been done). And these differences are not simply statistically significant, they are socially significant, as the large body of research on language in education shows (e.g. the pioneering study of Shirley Heath 1983, and the many studies done by Bernstein's research group, 1971, 1975, 1987). Bernstein is a sociologist, and he was not interested in merely describing linguistic differences. He wanted to embed them in a more general social theory in which one could see how differences in social class position led to differences in habits of language use, which in turn tended, in the context of a society and particularly an educational system shaped by those of the more powerful classes, to assign children of the less powerful classes to jobs and lives in which they would not wield power. Bernstein, too, sought to connect discourse to larger social relationships and processes. While his model emphasizes the reproduction of social relationships rather than social change and social dynamics, it otherwise represents an impressive general synthesis (see Bernstein 1981). And here once again we find an intermediate notion, code , or semantic orientation, that serves to bridge between the event and the larger social system. So far, all of these social theories of discourse have begun from the discourse side, that is from the text or utterance, and sought to explain its features by their social origins or functions. We turn now to two theorists who are mainly interested in the larger social system, but who have also introduced intermediate notions very similar to those of Bakhtin, Halliday, and Bernstein in order to connect that system to specific texts of discourse and action.  Foucault's Discursive Formations  Michel Foucault saw himself primarily as a historian: not as someone who sought to retell the past as it had been, but as someone who tried to describe how we today construct our continuities and discontinuities with many pasts. The texts and artifacts of the past are objects in our present-day world, and it is by way of our present-day notions of similarity and difference, continuity and discontinuity, that we construct their

historical meaning in the present day, and for the present day, by construing relationships among these objects and ourselves. For most historians, the primary objects of the past are texts, written documents surviving in various archives. Modern scholars have also learned to read painting and statuary, architecture and battlefields as texts as well. Historians, like anthropologists, are philologists; they need to find ways of reading texts even though they are not members of the communities that made these texts and in which the texts had their original meanings. Those original meanings are not recoverable; we can never know if we have reconstructed them or not. But we can still learn from them, learn from the ways in which they are different from the texts we make today. What is critical in this enterprise is how we put together different texts: which texts go together, and why, and how? A diary, a set of tax records, a chronicle, a taxonomy of diseases, a treatise on alchemy, a record of a trial for witchcraft. Foucault sought to build a general model of how our picture of the past, of our continuities and discontinuities with it, depends critically on our sense of the possible ways in which texts can be combined. He was building, in one sense, a general theory of intertextuality for the practice of history, and so in part for the practice of social science. Of all the theorists considered here he is the one most concerned with change. He reflexively situates his own discourse inside the systems he analyzes, at least insofar as he recognizes that how a historian looks at texts is itself part of a discursive formation built over historical time out of foundations that may include those same texts. When Foucault points out the ideological functioning of some way of speaking, he usually does so because, having seen its historical continuities and discontinuities with other ways of speaking, he can no longer regard it as a natural or inevitable product of common sense necessity (e.g. Foucault 1980). Foucault's major theoretical statement is The Archeology of Knowledge (1969), and this is usefully read in the context of his analysis of the emergence of social science discourse in the modern world (The Order of Things , 1966). Foucault's analysis of the principles of intertextuality is thoroughly postmodern (despite misunderstandings of it in the 1970s as a version of structuralism) and probably the most sophisticated possible in our time. Only its failure to engage with linguistic analyses of discourse limits its usefulness. It is not possible to know in terms of linguistic features of texts exactly how to interpret many of Foucault's theoretical principles, and while he sketches the general principles, there are no explicit examples to show us how to actually analyse the relations of specific texts. These are of course implicit in much of the rest of his work, but we have to recognize that his notion of a discursive formation cannot be equated exactly with any linguistically defined notion of a discourse formation . For our purposes, however, this is not necessary. We are interested in how Foucault uses the notion of the discursive formation to help bridge between texts and social systems, and we can see that functionally the discursive formation is an intermediate notion of exactly the same kind as the others we have identified. Here, for example, is Foucault's version of the principle of general intertextuality: "At its very root the statement (enoncee) has a dispersion over an enunciative field in which it has a place and a status, which arranges for it possible relations with the past and opens up possible futures. ... There is no free, neutral, independent statement; a

statement always belongs to a series or a whole, plays a role among other statements, is part of a network of statements. ...  There is no statement that does not presuppose others; that is not surrounded by a field of coexistences, effects of series and succession, a distribution of functions and roles. If one can speak of a statement as such, it is because a sentence or proposition figures at a definite point, with a specific position, in an enunciative network that extends beyond it." (Foucault 1969: 99). The enunciative field or network specifies, roughly, the rules of use of a statement in various contexts in relation to other statements. Another way of saying this is that statements tend to be used together in certain typical patterns (discursive practices) and to form systems (discursive formations) that relate statements to one another according to a variety of principles. "A discursive practice can be defined as ... a body of anonymous historical rules, always determined in the time and space that defined a given period, and [which determines] for a given social, economic, geographical, or linguistic area, the conditions of operation of the enunciative function." (1969: 117). A discursive formation for Foucault is defined by four kinds of relations among statements: those which determine what sorts of discursive objects (entities, topics, processes) the discourse can construct or talk about; those that specify who can say these things to whom in what contexts; those that define the relations of meaning among statements, including how they can be organized to form texts; and finally those that tell us what the alternative kinds of discourses are that can be formed in these ways and how they can be related to each other as being considered equivalent, incompatible, antithetical, etc. Foucault's notion of the discursive formation is thus more powerful than any of the notions we have encountered previously because it includes the rules for how these others are to be related to one another (i.e. what kinds of heteroglossic relations can exist among these narrower notions of discourse formations). Finally, we need to hear Foucault on discursive change: "A change in the order of discourse does not presuppose 'new ideas', a little invention and creativity, a new mentality, but transformations in a social practice, perhaps also in neighboring practices, and their mutual articulation. I have not denied the possibility of changing discourses: I have deprived the sovereignty of the subject of the exclusive and instantaneous right to it." (1969: 209). That is, discursive change is cultural change, it is systemic change. It is not the province of individual action, though it may originate in an individual event; it requires that a social community change its ways of speaking and doing. We see that for Foucault as well as Halliday the focus is on social practices, habits of activity characteristic of a community, not on individual acts of intentionality. For Foucault, the discursive formations that tell us what people are saying and doing in a historical period are systems of doings, not of doers as such. Foucault provides a discussion of what he calls the "subject-positions" defined by a discourse formation, the social roles of the speakers of these discourses. He seems to suggest to many people that

we can use the notions of discourse formations to define individual subjects insofar as they are participants in a discourse. We will return in Chapter 5 to this complex question. For now the important point is that one can give, as Foucault does, very complex and subtle accounts of social relationships and their historical changes in terms of discourse formations. And by reading Foucault against the background of Bakhtin, Halliday, and Bernstein, we can see once again how intermediate notions of a particular kind help to connect texts or events and the social systems in which such texts can occur, do occur, and make sense. If there is one element of this synthesis that is still rather weak, it is the problem of how to relate discourse formations seen in overview as characteristic of societies and their cultures with the actual lives of individual people who enact these discourses, and in enacting, potentially change them. Bernstein has begun to give an account of this process: how we are each socialized into the discourse patterns and habits, the coding orientations, characteristic of our social class, gender, subculture, etc. But the theorist who has made the fullest effort to provide a general theory of how people of different social categories acquire their social habits is probably Pierre Bourdieu.  Bourdieu and Discourse Habitus  By now it should be getting pretty clear just what all these intermediate notions which bridge between texts or events on the one hand and larger social systems characteristic of whole communities, on the other, have in common. They are all notions of what is typical in a community: typical habits and patterns of discourse and action. Every text or event is unique, but it can also be seen as an instance of some kind or type of text or event that recurs in a community and is recognizable as such. Most general are what we might call activity formations , the typical doings of a community which are repeatable, repeated, and recognized as being of the same type from one instance or occurence to another. A baseball game, a train ride, writing a check, making a phone call. We could also call these action genres . Among the special cases of action genres are speech genres and written genres , but these are clearly also definable as the products of the activities that produce them. Genres are rather specific in their properties, having definite beginnings, ends, and stages along the way. Notions like register and discourse formation can be made a bit more general to handle kinds of language apart from such neat packaging, and so also can activity formations. What is important here are the relations between text or event and formation or genre, on the one hand, and those between formations or genres and larger issues of social structure and process on the other. Every text or event takes its meaning in part from being seen in the community as an instance of one or more formations. We interpret it against the background of other instances of the same formations to see how it is distinctive, and we contrast it with instances of other formations. But different formations (codes, genres, registers, voices of heteroglossia, discursive formations) are not just different: they have systematic relations to one another, and those relations define and are defined by the larger social relationships of classes, genders, age groups, political constituencies, and significant social divisions of every kind. The model is recursive: each level is defined by its relations to the other levels in the model. So, for

instance, social class is defined by the fact that not all activities in the community are equally likely to be practiced by all people. People are defined by the activities they participate in, and significant social categories of people by the intersections of groups of related activities, including the discourse practices by which we label people as members of social categories. Models of this degree of complexity and recursiveness appear to be necessary when dealing with human social systems (see the final section of this book, "Making Trouble", and also Chapter 6 for some of the reasons why, and for a fuller discussion of how the various levels of the model integrate with one another). A social theory which has the requisite degree of complexity (except perhaps for underplaying the role of discourse and the inherent dynamic features of the system that lead to its continual changes) is Bourdieu's (1972, 1990) theory of social habitus . Bourdieu has made some special contributions to social models. One of these is his efforts to link social abstractions like the habits, attitudes, preferences, dispositions, and actions characteristic of a social class, gender, age group, etc., to the actual life-trajectories of bodily persons.  Bourdieu has noticed, as have many others trained as he was in social anthropology, that members of different cultures not only talk differently (using different languages, discourse formations, coding orientations), but they even walk differently. They carry themselves differently, with a body hexis distinctive to their culture (and gender, and age group, etc.). This suggested to Bourdieu that cultural and subcultural dispositions of all kinds are literally embodied in people. Bourdieu here rejects the great Cartesian split which seems so clearly to function ideologically in the discourses of the human sciences. He takes something usually thought of as belonging to the domain of "mind": how we perceive things, how we feel about them and react to them, our habits and preferences and attitudes and dispositions to action (including to discourse) and makes them matters of body . By the same move, he renders unnecessary the dichotomy between matters characteristic of groups, communities, social categories like gender and age, etc. and matters characteristic of individuals. He speaks of culture as directly embodied in persons. Persons with such dispositions to action embodied in them tend to act in ways that reinforce these dispositions, or in many cases complementary dispositions, in others. Thus social relationships also become embodied. Cultural habitus for Bourdieu is an embodied system of sociologically structured and structuring dispositions. We acquire these dispositions in the course of living our lives, interacting with the social and material (especially the human- made) environment, which consists of other people acting out of these dispositions and the material effects of such actions in the world. We do not all acquire the same dispositions of course, for we live different lives, have different characteristic experiences, participate in different activities with different frequencies, and occupy different roles in the activities in which we do participate. The dispositions of the habitus are more alike for those who lead more similar lives, and progressively become less alike for those who typically engage in different roles and different activities. Habitus can be as specific as the dispositions acquired by a trained athlete or dancer, dispositions specific to their sport or their style of dance training. Or it can be as general as the dispositions that distinguish males and females, or workers and managers.

 Bourdieu's other special contribution is his emphasis on the distinction between synoptic and participatory views of human activity. Synoptic views stand outside of the process of enactment of an activity, generally describing it after it is finished, or as an ideal formula that applies to the typical case. Participatory views (I also use the term dynamic view in this sense) look at human action from the viewpoint, not of an outside observer, but of a participant, for whom every aspect of the on-going action is contingent, dependent on the next move, the next response or reaction, and so on the various strategies by which we get through the activity, bringing it to some sort of, usually conventional, conclusion. The notion of habitus or embodied cultural disposition also links these two perspectives together. The habitus is what shapes our responses to the myriad unpredictable contingencies of the moment, and shapes them in such a way that, on the whole, when the synoptic accounts are totaled up, things have turned out in the way typical of goings-on in our community. The habitus mediates between a synoptic view of activity formations characteristic of a community and a dynamic view of the processes by which these activities are actually enacted on specific occasions by human actors. We should not be surprised then that Bourdieu has extended Bakhtin's notions of heteroglossia in much the way that I have in my own work (cf. Bourdieu 1991, Lemke 1988c and next chapter). He regards the social relationships among discourse voices as being structured by, and in turn contributing to the structuring of, the social relationships of power among different positions in the social field (defined by social class, by age, by gender, etc.).  Bourdieu's basic metaphor for this is an economic one. He construes an economics of linguistic transactions in which utterances or discourses are the products which producers offer on a market to potential consumers. Each producer, by virtue of membership in a subcommunity or position in the larger web of social relations (what I will sometimes call a caste or subcaste), has some linguistic habitus, some embodied system of dispositions to speak in a particular way. The link to material embodiment is most evident in the case of social "accent" or norms of pronunciation, and from there to social dialects (including lexicon and grammar), semantic coding orientations (cf. Bernstein and Hasan references above), and even genres and discourse formations, is a reasonable progression. So a speaker speaks partly in ways typical of his or her social position and caste membership: in accent, in grammar and lexis, in semantic dispositions, and in likelihood of using particular genres and registers to produce discourses of recognizable types with definite viewpoints on their subjects. Bourdieu also recognizes that this process of discourse production maps forms and contents onto one another, so that in the finished product we can no longer distinguish them. In fact, in many cases there are no intertexts, no alternatives available in which we could see the same content in a different form, or vice versa, since we cannot in general find just any social or political point-of-view combined with any statement about the world. Bourdieu provides us with a way of connecting the relations among the contents and viewpoints of various discourse formations, or social voices of heteroglossia, with the relations among the social positions of their authors. Since his view of linguistic habitus includes, as it should, interpretive, or consumer, dispositions toward discourse as well

as producer dispositions, he can also show us that we all evaluate the worth of discourses, and even of utterances, from our own social viewpoints. These evaluations are part of the meaning a linguistic act or text has for us, a critical part of its/our textual politics. We evaluate some accents as better or more prestigious than others, some dialects as better, some realizations of the norms of a discourse type as better. In doing so, we hear and read (and ourselves produce) language always against the social background of these evaluations. While evaluations may differ from one caste to another in a society, there are generally dominant norms of evaluation, which are those of the dominant caste, and which are to some extent accepted as natural by members of other castes. In any case, everyone knows up to a point what these dominant norms are and speaks and evaluates at least in relation to them if not always strictly according to them. They are facts of social life. And they are what they are because of the overall power of the dominant caste to maintain their dominance in discourse as in all else. Bourdieu takes this so much for granted, referring only to the "field of power," as the social background for these relations among discourses and their evaluations, that we do not seem to get in most of his work any very explicit grounding of textual politics in the politics of coercive power. Bourdieu's view of power is multiplex: there is economic power, symbolic power of many kinds, social influence -- each grounded in its own sort of capital and in caste-specific dispositions to acquire and use that capital. But like most of us, Bourdieu looks rather little at the most primitive forms of capital: physical strength, weapons and the dispositions to use them to control the behavior of others. Because there are so many other ways in which social control of behavior is exercised in modern society, we prefer to overlook the most basic one, and so we may miss the important ways in which coercive power grounds the efficacy of all other forms of power, and the role of bodily materiality in this as well. Bourdieu's sociology seems generally well suited to help us bridge from particular texts and events to larger macrosocial structural relations, particularly from the intermediate formations themselves (of which Bourdieu has relatively little to say) to, on the one hand, the materiality of situations and human participants (by way of the embodied dispositions of the caste-specific habitus), and on the other, the relationships of social power among significant social groups. But it is inevitable, in this picture of the relationships between discourses and the social positions of their authors, that Bourdieu's discourse also is limited by his social position. I have already suggested that as an upper-middle class intellectual, he may be overinclined to emphasize the role of symbolic capital and less disposed to focus on that of coercive power.  Whether it can be laid to his own social positioning or not, we should also be aware that Bourdieu's view of social processes, however dynamically he sees the constitution of social structure, remains basically a static view. He is not concerned primarily with longterm historical change, or indeed with radical and revolutionary change. It is indeed hard to imagine anything but the most gradual and piecemeal changes in social life as Bourdieu describes its basic mechanisms. I believe as well that his discourse embodies a masculinist disposition, which, while very sensitive to the general social domination of females by males, still tends to see all of social life as a competitive struggle for profit and distinction in a way particularly characteristic of masculinist perspectives in our culture (see also Lemke 1993c). And it is not surprising either that little attention is

given to the viewpoints on social life of the very young or the very old, currently still the most invisible of the basic biases of our intellectual culture. The social theories of discourse presented in this chapter fit together like the pieces of a puzzle. They develop essentially similar approaches to the roles of intertextuality, cultural formations, and the web of social relationships in the discursive construction of meaning. Taken together, they also point to additional factors which need to be better theorized: the materiality of meaning-making processes, the discursive construction of individuality and subjectivity, the role of coercive power in the social order, the politics of our own theories.  ? Selected References For complete references see: http://www-personal.umich.edu/~jaylemke/tp-biblio.htm .I have marked in RED the items that I would consider key to our discussions.

Bakhtin, M. (aka V.N. Voloshinov) 1929. Marxism and the philosophy of language. 1986 edition. Cambridge, MA: Harvard University Press.

--- 1935. Discourse in the novel. In M. Holquist, Ed. The Dialogic Imagination (1981). Austin TX: University of Texas Press.

--- 1953. Speech Genres and Other Late Essays. 1986 edition. Austin: University of Texas Press.

Bernstein, B. 1971. Class, Codes, and Control. Volume 1. London: Routledge.

--- 1973. Class, Codes, and Control. Volume 2. London: Routledge.

--- 1981. Codes, modalities, and the process of cultural reproduction. Language in Society 10 (3): 327-364.

--- 1987. "Social class, codes, and communication." In Sociolinguistics: An International Handbook of the Science of Society, U. Ammon, N. Dittmar, & K.J. Matthier, Eds. Berlin: de Gruyter.

Bourdieu, P. 1972. Outline of a Theory of Practice. Cambridge: Cambridge University Press.

--- 1979. Distinction: A Social Critique of the Judgement of Taste. Cambridge, MA: Harvard University Press (1984 edition).

--- 1990. The Logic of Practice. Stanford CA: Stanford University Press.

--- 1990. Homo Academicus. Stanford: Stanford University Press, 1990.

--- 1991. Language and Symbolic Power. Cambridge, MA: Harvard University Press.

Bourdieu, P. & Passeron, J.C. 1977. Reproduction in Education, Society, and Culture. Beverly Hills, CA: Sage.

Bourdieu, P. & L.J.D. Wacquant. 1992. An Invitation to Reflexive Sociology. Chicago: University of Chicago Press

Foucault, M. 1966. The Order of Things. New York: Random House.

--- 1969. The Archeology of Knowledge. New York: Random House.

--- 1980. The History of Sexuality. Volume 1. New York: Random House.

Gee, J.P. 1990. Social Linguistics and Literacies. London: Falmer Press.

Halliday, M.A.K. 1967. Grammar, Society, and the Noun. London: H.K. Lewis for University College London.

--- 1975. Learning How to Mean. London: Edward Arnold.

--- 1976. In G. Kress, Ed., Halliday: System and Function in Language. London: Oxford University Press.

--- 1977. Text as semantic choice in social context. In van Dijk, Teun A. and Pet?fi, J. (Eds.) Grammars and Descriptions. Berlin: de Gruyter.

--- 1978. Language as Social Semiotic. London: Edward Arnold.

--- 1982. The de-automatization of grammar. In J. Anderson, Ed., Language Form and Linguistic Variation, pp.129-159. Amsterdam: John Benjamins.

--- 1985b. Spoken and Written Language. Oxford: Oxford University Press. (2nd edition, 1989).

--- 1988. On the language of physical science. In M. Ghadessy, Ed. Registers of Written English. London: Pinter Publishers.

--- 1990. New ways of meaning. In S. Efstathiadis, Ed. Selected Papers from the Ninth World Congress of Applied Linguistics, special issue of Journal of Applied Linguistics, v.6, (Thessaloniki: Applied Linguistics Association of Greece). Re-printed in Language in a Changing World, Applied Linguistics Association of Australia (Occasional paper No. 13), 1993.

--- 1991. Towards probabilistic interpretations. In E. Ventola, Ed., Recent Systemic and Other Functional Views on Language. (pp. 39-62). Berlin-New York: Mouton/deGruyter.

--- 1992. How do you mean? In M. Davies and L. Ravelli, Eds., Advances in Systemic Linguistics: Recent Theory and Practice. [pp.20-36]. London: Pinter.

--- 1993. Towards a language-based theory of learning. Linguistics and Education 5(2): 93-116.

--- 1994. An Introduction to Functional Grammar. London: Edward Arnold. 2nd edition.

Halliday, M.A.K. & Fawcett, R. 1987. New Developments in Systemic Linguistics. London: Frances Pinter.

Halliday, M.A.K. & Hasan, R. 1976. Cohesion in English. London: Longman.

--- 1989. Language, Context, and Text. London: Oxford University Press.

Halliday, M.A.K. & Martin, J.R. 1993. Writing Science. Pittsburgh: University of Pittsburgh Press.

Hasan, R. 1984a. "Coherence and Cohesive Harmony." Understanding Reading Comprehension. Ed. J. Flood. Newark, DE: International Reading Association.

--- 1984b. "The structure of the nursery tale." In L. Coveri, Ed., Linguistica Testuale. Rome: Bulzoni.

--- 1985. Lending and borrowing: From grammar to lexis. In J. Clark, Ed., The Cultivated Australian (Beitrage zur Phonetik und Linguistik, 48), pp.55-67. Hamburg: Helmut Buske Verlag.

--- 1986a. The grammarian's dream: Lexis as most delicate grammar. In M.A.K. Halliday and R. Fawcett, Eds., New Developments in Systemic Linguistics. London: Frances Pinter.

--- 1986b. The ontogenesis of ideology. In T. Threadgold, E.A.Grosz, Gunther Kress, & M.A.K. Halliday, Eds., Semiotics, Ideology, Language. Sydney: Sydney Association for Studies in Society and Culture.

--- 1988. The analysis of one poem. In L.M. O'Toole & D. Birch, Eds., Functions of Style. London: Pinter.

--- 1989a. The structure of a text. In Halliday, M.A.K. & Hasan, R. Language, Context, and Text. London: Oxford University Press.

--- 1989b. Semantic variation and sociolinguistics. Australian Journal of Linguistics, 9, 221-76.

--- 1994. "The conception of context in text." In M. Gregory & P. Fries, Eds. Discourse in Society: Functional Perspectives. Norwood, NJ: Ablex Publishing.

Hasan, R. & Cloran, C. 1990. A sociolinguistic interpretation of everyday talk between mothers and children. In Halliday, M.A.K., John Gibbons, and Howard Nicholas, Eds., Learning, Keeping, and Using Language. Amsterdam: John Benjamins.

1996. Hasan, Ruqaiya. Ways of Saying: Ways of Meaning: Selected Papers of Ruqaiya Hasan. (Edited by Carmel Cloran, David Butt and Geoffrey Williams). London/ New York: Cassell.

Kristeva, J. 1980. Desire in Language. New York: Columbia University Press.

Lemke, J.L. 1983a. Thematic Analysis: Systems, Structures, and Strategies. Semiotic Inquiry 3(2): 159-187.

--- 1985. Ideology, intertextuality, and the notion of register. In Benson, J.D. and Greaves, W.S. (Eds.), Systemic Perspectives on Discourse. Norwood, NJ: Ablex.

--- 1988c. Discourses in conflict: Heteroglossia and text semantics. In Benson, J. D. and Greaves, W. S. (Eds.), Functional Perspectives on Discourse. Norwood, NJ: Ablex Publishing.

--- 1989b. "Social semiotics: A new model for literacy education" In D. Bloome, Ed. Classrooms and Literacy (pp. 289-309). Norwood, NJ: Ablex Publishing. 1989.

--- 1989c. Semantics and social values. WORD 40(1-2): 37-50.

--- 1990a. Talking Science: Language, Learning, and Values. Norwood, NJ: Ablex Publishing.

--- 1990b. Technical discourse and technocratic ideology. In M. A. K. Halliday, John Gibbons, and Howard Nicholas, Eds., Learning, Keeping, and Using Language: Selected Papers from the 8th AILA World Congress of Applied Linguistics. Vol. II. pp. 435-460. Amsterdam: John Benjamins. 1990.

--- 1992a. Interpersonal Meaning in Discourse: Value Orientations. In M. Davies and L. Ravelli, Eds. Advances in Systemic Linguistics: Recent Theory and Practice. [pp.82-104]. London: Pinter. 1992.

--- 1993c. "Practice, Politique, Postmodernism." A review of Pierre Bourdieu and Lois J.D. Wacquant's _An Invitation to Reflexive Sociology_. Postmodern Culture [PMC-LIST on Listserv@listserv.ncsu.edu. Get PMC-LIST REVIEW-4.993]. 1993.

--- 1993d. "Intertextuality and Educational Research." Linguistics and Education 4(3-4): 257-268, 1993.

--- 1994a. "Discourse, Dynamics, and Social Change." Cultural Dynamics 6(1): 243-275. [Special issue, Language as Cultural Dynamic, M.A.K. Halliday, Issue Editor].

--- 1994b. Intertextuality and text semantics. In M. Gregory and P. Fries, Eds. Discourse in Society: Functional Perspectives. Norwood, NJ: Ablex Publishing.

--- 1995. Textual Politics: Discourse and Social Dynamics. New York/London: Taylor & Francis.

Ure, J. & Ellis, J. 1974. Register in descriptive linguistics and linguistic sociology. In O. Uribe-Villegas, Ed., Issues in Sociolinguistics. The Hague: Mouton.

Voloshinov, V.N. 1929. Marxism and the Philosophy of Language. Cambridge, MA: Harvard University Press (1986 edition).

Postscript:

In addition to these references, most of which are prior to 1995 (a few have been updated here), one could add some examples of the work of scholars in the Critical Discourse Analysis movement (in which I have peripherally participated), including:

Fairclough, Norman:

Language and Power, London: Longman 1989 (second revised edition 2001)

Discourse and Social Change, Cambridge: Polity Press 1992 (Portuguese translation, 2001; Chinese translation in preparation)

Critical Discourse Analysis, London: Longman 1995

Discourse in Late Modernity - Rethinking Critical Discourse Analysis, Edinburgh: Edinburgh University Press 1999 (with Lilie Chouliaraki)

Van Dijk, Teun:

Discourse and Discrimination (Detroit: Wayne State U.P, 1988)(with Geneva Smitherman, Eds.). Racism and the Press (London: Routledge, 1991) Elite Discourse and Racism (Newbury Park, CA: Sage, 1993). Discourse Studies, 2 vols. (Ed.). (London: Sage, 1997). Ideology (London: Sage, 1998). Racism at the Top (Klagenfurt, Drava Verlag, 2000)(with Ruth Wodak, Eds.).

Wodak, Ruth:

The Discursive Construction of National Identity. Edinburgh: University Press 1999 (with R.de Cillia, M.Reisigl, K.Liebhart) Racism at the Top. Klagenfurt: Drava, 2000, (with Teun A. van Dijk). Methods of Text and Discourse Analysis. London: Sage, 2000, (with M.Meyer, St.Titscher, E. Vetter) Methods of Critical Discourse Analysis. London: Sage, (forthcoming) (with M.Meyer) Discourse and Discrimination.  Rhetorics of Racism and Antisemitism. (with Martin Reisigl), 2000, London: Routledge 

EU Discourses on Un/employment, An interdisciplinary approach to employment policy-making and organizational change. (with P.Muntigl, G.Weiss), 2000, Amsterdam: Benjamins .

Theory of Complex Self-organizing Systems

Traditional European philosophy has favored theories in which causality is the dominant explanatory principle: things happen because someone, or by extension something, makes them happen. It is clear that this reflects a fantasy, probably particularly masculinist, of power over life, the world, and especially other humans (desirable females, dangerous males in the dominant view). But causality has rarely provided adequate accounts of most systems because they consist of many interacting parts and the behavior of the system as a whole, and often of the individual parts, is an complex aggregation of the interactions of all the parts -- and no part controls the whole, or can even control another part outside the influence of the rest of the system. Such systems are said to be 'self-organizing' and the behavior of aggregates of components is said to be 'emergent'. In these systems, which certainly include living organisms, ecosystems, and social or ecosocial systems, there are no isolated controlling agencies. There is no all-powerful father, boss, or king. There are no control hierarchies among components: no generals, captains, and soldiers. Self-organizing systems are inherently democratic, and eurocultural philosophies basically are not.

The modern theory of self-organization phenomena has several roots: cybernetics, which did look for control hierarchies, but quickly saw more complex behavior; organismic biology, especially the early traditions of "holism" that opposed the "reductionism" of physics; ecosystem theory; the autopoesis theory of Varela and Maturana; auto-catalytic and cross-catalytic reaction theory in chemistry; mathematical ecology; cellular automata theory; thermodynamics and statistical physics of irreversible processes; and the mathematics and physics of non-linear equations.

For an overview in relation to social dynamics, see: Lemke, Textual Politics, chapter 6; Lemke, Cultural Dynamics article; and Lemke, Downward Causation article. References in these works cite the major sources of the theory. Key names and associations are:

Stuart Kauffman -- cellular automata, evolutionary and developmental biology Stanley Salthe -- developmental and evolutionary biology, origins of life,

hierarchical levels in complex system Ilya Prigogine -- thermodynamics and statistical physics of irreversible process;

origins of complexity Gregory Bateson -- cybernetic models of ecosocial interactions, meta-learning Francisco Varela and Umberto Maturana -- autopoetic systems theory Howard Odum -- systems ecology H. Ross Ashby -- cybernetics

Also: Norbert Wiener, J. Doyne Farmer, H. Poincare, L. van Bertalanffy, R. Thom, N. Luhmann

General Social Theory

All events in an ecosocial system take place in a sociological and cultural context as well as an ecological context. A general theory of social relations is a necessary part of an ecosocial-semiotic theory. No adequate theory exists, primarily because it is not possible to construct such a theory from inside the system: all social theorists are enmeshed in the social relations they describe; we cannot escape being part of a gender system, a class system, a specific cultural-historical epoch, etc. But some theories are more useful than others (see Lemke, Textual Politics, chapter 2 for criteria).

A general social theory usually identifies the categories of persons constructed by the practices of a community, and specifies the relations among these categories in terms of power, prestige, and specialized function within the community. It then tries to explain how these categories and relationships function, how they came to be established historically, what keeps them going, and how they are always changing. An ideal theory relates the global-scale phenomena of the community to the local-event scale phenomena (e.g. gender and class to conversation and action). All existing macrosocial categories are subject to radical critique and should not be taken as absolute, merely as indications of the existence of macrosocial phenomena.

The principal, useful social theories and their basic principles and formulators are:

neo-Marxism: class-based analysis; K. Marx, L. Althusser, A. Gramsci M. Foucault: discursive formations, archeology, power relations in historical

context P. Bourdieu: habitus, structuration, capitals, fields; comprehensive except for

historicity B. Latour: actant networks, heterogeneity of types and scales; very sophisticated N. Luhmann: systems theory, discourse; more cybernetic than self-

organizational C. Geertz: thick descriptions, hermeneutic ethnography; phenomenological

approach Ethnomethodology: primarily a microsocial theory, but very useful, also with

roots in phenomenological philosophy (E. Husserl); A. Schutz, H. Sacks, A. Cicourel, E. Schegloff, and related work by E. Goffman

Also: A. Giddens, B. Bernstein, R. Harre

General Semiotics

Semiotics is the theory of the production and interpretation of meaning. It's basic principle is that meaning is made by the deployment of acts and objects which function as "signs" in relation to other signs. Systems of signs are constituted by the complex meaning-relations that can exist between one sign and another, primarily relations of contrast and superordination/subordination (e.g. class/member, whole/part). Signs are deployed in space and time to produce "texts", whose meanings are construed by the mutually contextualizing relations among their signs.

There are two major traditions in European semiotics: F. de Saussure (Swiss-French), semiology; and C.S. Peirce (Anglo-American), semiotics. Saussure's approach was a

generalization of formal, structuralist linguistics; Peirce's was an extension of reasoning and logic in the natural sciences.

Subsequent theorists of importance include: L. Hjelmslev, R. Barthes, G. Bateson, J. Lacan, S. Freud, B.L. Whorf, B. Malinowski

Also: U. Eco, J. Baudrillard, M. Serres, J. Lemke, M.A.K. Halliday, T. van Leeuwen, N. Goodman

Social Semiotics

General Semiotics tends to be formalistic, abstracting signs from the contexts of use; Social Semiotics takes the meaning-making process, "semiosis", to be more fundamental than the system of meaning-relations among signs, which are considered only the resources to be deployed in making meaning. Social semiotics examines semiotic practices, specific to a culture and community, for the making of various kinds of texts and meanings in various situational contexts and contexts of culturally meaningful activity. Social semiotics therefore makes no radical separation between theoretical and applied semiotics and is more closely associated with discourse analysis, multimedia analysis, educational research, cultural anthropology, political sociology, etc.

Principal developers of Social Semiotics include: M.A.K. Halliday, R. Hasan, J.R. Martin, J.L. Lemke, G. Kress, R. Hodge, T. van Leeuwen, P. Thibault, T. Threadgold, L.M. O'Toole

Multimedia Semiotics

Multimedia semiotics is based on the principle that all meaning-making, because it is a material process as well as a semiotic practice, necessarily overflows the analytical boundaries between distinct, idealized semiotic resource systems such as language, gesture, depiction, action, etc. Every material act and sign can be, and usually is, construed in relation to more than one system of sign relations (e.g. a written word is both a linguistic sign and a visual orthographic one; a spoken word is also construed in relation to its non-linguistic acoustical qualities; an image is interpreted both visually and usually also linguistically; etc.). Therefore it becomes important to study how different sign-systems are physically and semiotically integrated in texts and multimedia productions of various kinds.

One approach is to analyze various sign-systems separately and then study their integration, particularly emphasized are linguistic, gestural, graphical-pictorial, musical, and motor-actional semiotic systems. The greatest progress has been made for visual semiotics, especially in the work of: T. van Leeuwen, G. Kress, L.M. O'Toole, E. Tufte, S. Eisenstein, R. Arnheim, R. Barthes, C. Metz, J. Bertin

A second approach is to attempt to directly analyze the interconnections among signs belonging to different systems as they are deployed in particular texts, e.g. Lemke, Multiplying Meaning.

Functional and Systemic-Functional Linguistics

Functional linguistics opposes itself to formal linguistics in much the same way that social semiotics contrasts with general or formal semiotics. Functional linguistics looks at language as as system of resources shaped by the functional uses of these resources in particular cultures over history. Its unit of analysis is text-in-context. It seeks to describe and explain how particular linguistic units and formal relationships are deployed to make possible the construal of particular meanings.

There are many traditions of functional linguistics. A simplified story is that they originated in Eastern Europe where Russian formalism was always more concrete and sociologically aware of context than was the West European variety. From this eastern formalism came the Moscow school functionalism (Propp, Voloshinov, Bakhtin, R. Jakobson), and the Prague School functionalists (V. Mathesius, J. Mukarovsky, F. Danes), which then migrated by way of B. Malinowski to England, where it developed into British functionalism (J.R. Firth, M.A.K. Halliday, J. McH. Sinclair), and also via Germany to the U.S. (E. Sapir, B.L. Whorf, M. Silverstein, J. Gumperz) where it flourished more in anthropology than in linguistics itself. The English branch came via Halliday and his partner R. Hasan to Australia, and with M. Gregory to Canada. J.R. Martin studied with Gregory and Halliday and also emigrated to Australia. I met Halliday and worked with him, Hasan, and Martin in Sydney, where there developed additionally the so-called Newtown Semiotic Circle, including T. Threadgold, G. Kress, P. Thibault, and others, which developed the Social Semiotics model as a generalization of Halliday's social linguistics. Other branches of functionalism are Danish (L. Hjlemslev), French (A. Martinet, C. Hagege), Dutch (T. van Dijk), and German-Austrian (W. Dressler). There are close associations between functionalism and sociolinguistics (e.g. W. Labov, J. Gumperz, J. Fishman, W. Dressler) in many traditions, as between functional linguistics and anthropological linguistics.

Systemic-functional linguistics (SFL) is the variant developed by Michael A.K. Halliday and is probably the most fully elaborated and useful system for discourse analysis and various areas of applied linguistics. It has more distant affinities to tagmemic linguistics (K. Pike) and stratificational linguistics (H.A. Gleason). The newest variant is a computational linguistics deriving directly from Halliday, developed mostly by C. Matthiessen and his colleagues in a number of projects. For more information on SFL, see the Systemics Website.

Discourse Analysis

Because the most developed branch of semiotics is the study of language signs and their use, it is possible to study the sign relations within (discourse semantics) and between (intertextuality) linguistic texts in great detail; this is very useful as a beginning in the study of other phenomena.

For a general overview of discourse analysis methods, see Lemke, Analyzing Verbal Data.

The origins of discourse analysis lie in classical Rhetorical Theory (Aristotle, Cicero, Longinus) and its successors. Modern discourse analysis derives initially from the formalist, structuralist work of V. Propp on the morphology of Russian folktales, the forerunner of Genre Theory. It was further developed by the Russian and Prague schools of functional text analysis, both for literary and nonliterary genres of writing,

and to a small extent also for speech (e.g. J. Mukarovsky, R. Jakobson, V.N. Voloshinov, M.M. Bakhtin, V. Mathesius).

Recent developments in discourse analysis are based on the ordinary language philosophy movement traced to L. Wittgenstein (also: J.L. Austin, J.R. Searle, H.P. Grice) -- not all of which is reliable, however -- and on the functional linguistic traditions in the U.S. (E. Sapir, B.L. Whorf, M. Silverstein, J. Gumperz) and in England (J.R. Firth, M.A.K. Halliday, J. McH. Sinclair), Canada (M. Gregory, English), and Australia (J.R. Martin, Canadian; R. Hasan, Pakistani who studied in England).

The major subsections of Discourse Analysis theory within the Systemic-Functional and Social Semiotics traditions, and the key developers are:

Register theory: variation of text with context; Halliday, Gregory, Hasan, Martin

Genre theory: specialized text-types; Propp, Bakhtin, Hasan, Martin Intertextuality: text-text relations; Bakhtin, J. Kristeva, R. Barthes, Lemke,

Thibault Text semantics: meaning made beyond the sentence; Halliday, Gregory,

Lemke, Martin Discursive formations: Foucault Thematic analysis: Lemke, Hasan, Halliday Cohesion analysis: Hasan, Halliday Speech act analysis: Austin, Hasan, Martin,

Thibault Attitudinal-Evaluative meaning: Bakhtin, Lemke, Martin, Thibault Conversation analysis: Sacks, Schegloff, S. Eggins, E. Ochs, J. Bruner Narrative analysis: Propp, Mathesius, Bakhtin, A. Greimas, W. Labov Theme-Rheme analysis: Halliday, P. Fries

Postmodernism and Critical Theory

Postmodernism and Critical Theory are broad rubrics for intellectual movements rather than specific theories, but they are essential parts of social semiotic analysis. Postmodernism derives from Post-Structuralism and Deconstructionism, which were initially criticisms of the Structuralist movement of the 1960s. Critical theory derives from neo-Marxism and Feminist theory, extended to include Post-colonial theory and Queer theory.

Structuralism was an off-shoot of general semiotics and formal linguistics and proposed that there were systematic abstract relations-of-relations among the many sign-elements of human culture, whether in kinship relations and village planning (Levi-Strauss), cognitive developmental operations (Piaget), or linguistic phonology (R. Jakobson). These patterns of relations could often be expressed mathematically in terms of abstract group theory and other algebraic structures. It was employed in part also by Barthes (literary semiology) and Lacan (psychodynamics). This view, while basically correct, was too narrow and was criticized for being too static and synchronic (Bourdieu), too ahistorical (Foucault), and too definitive (J. Derrida). Derrida in particular mounted a radical philosophical critique in which he pointed out that the very act of meaning making always presupposes an unanalyzable ground of the possibility of meaning and of sign systems, and that the dialectic of sign and ground produces an inherent instability or indefiniteness in any meaning. These were the Post-Structuralist and

Deconstructionist critics. Post-structuralism was very quickly succeeded by a more similar but more general critique of the master cultural narratives, metaphors, and assumptions of European (and European-American) culture (F. Lyotard), including those underlying the previous critical tradition of Marxism and neo-Marxism. This was and is Post-Modernism, though the term is already somewhat unfashionable in France. Post-Modernism made common cause with other new critiques, from non-European cultures, especially post-colonial ones, feminist critiques, and later gay and Queer Theory critiques. Some of these movements also allied themselves with neo-Marxism, as Critical Theory. The present rubric for Postmodern and Critical Theory work is often called Culture Studies.

Perhaps the most characteristic tenet of postmodern critical work is that everything that European philosophy and science has held to be fundamentally true at an abstract or programmatic level (ontology, epistemology, metaphysics, logic) is in fact a contingent, historically specific cultural construction, which has often served the covert function of empowering members of a dominant social caste at the expense of Others. It dismantles the most foundational procedures and assumptions whereby prior European philosophical traditions sought to establish universal truths or principles. It is fundamentally a revolutionary political movement, argued in intellectual terms. For a rather casual introduction to some of these issues, see Lemke, "Semiotics and the Deconstruction of Conceptual Learning."

Some useful sources: M.Foucault, J.-F. Lyotard, J. Baudrillard, M. deCerteau, G. Deleuze, B. Latour, D. Haraway, J. Butler, M. Serres, F. Jameson, H.K. Bhabha

Activity Theory and Actant-Network Theory

Relatively newer additions to the ecosocial dynamics toolkit are Activity Theory, associated with the work on A.N. Leontiev, a follower of Vygotsky, and the much more recent Actant-Network Theory (aka Actor-Network Theory), which derives mainly from the work of Bruno Latour. Familiarly known as AT and ANT.

Activity Theory essentially begins as a social-ecological psychology in which the unit of analysis is human activity. An activity derives its unity from the fact that all actions belonging to it tend toward some object/objective or goal. The notion of an activity is similar to activity-type in social semiotics or action genre in other terminology. While social semiotics does not accept intentional goals as a valid principle for defining actional units, it does agree with the developments of AT in the direction of understanding semiotic mediation (signs) in the same terms as artifact mediation (tools) in human activity. The later version of the theory has been developed particularly by Y. Engestrom. Vygotsky emphasized originally the role of discourse in mediating action. Ecosocial dynamics tends to see the unity of action as deriving mainly from habits at the level of the individual, and these in turn from actional social formations or socioculturally recognized activity-types. The views however are compatible, especially if one allows in AT the notion that goals are retrospective constructions and that the actual material objects of our actions are emergent in the actor-object-context dynamical system. (See for example the analysis in Lemke, Emergent Agendas paper.)

Actant-Network Theory has its origins in studies of the networks of interdependent social practices that constitute work in science and technology. Latour recognized that

semiotically both human actors and nonhuman participants (whether artifacts or naturalized constructs like bacteria) were equally actants in the sense of Greimas' narrative semiotics: they were defined by how they acted and were acted on in the networks of practices. The important fact here is not that humans and nonhumans are treated symmetrically (a given in social semiotics and ecosocial dynamics) but that they are defined relationally as arguments or functors in the network, and not otherwise. This leads to a relational epistemology which rejects the naive positivist view of objects or actors as existing in themselves prior to any participation in ecosocial and semiotic networks of interactions (including the interactions by which they are observed, named, etc.). Actantial-relational epistemology is not nominalism, but far more sophisticated. ANT has much in common with Ecosocial Dynamics, but adds one crucial observation: that the usual view of dynamical systems assumes that they have a local topology, and so events nearby in space and time are more relevant than those at a distance, leading to neat separation of scales of processes. ANT notes that the topology of networks is in general non-local, and further that semiotic artifacts are often the 'boundary objects' that mediate non-local, scale-breaking interconnections. This leads to a powerful generalization of ecosocial systems theory to include network topologies (and the rarer laminar topologies) and makes possible a general inquiry into scale-respecting vs. scale-breaking dynamics. See discussion in Lemke, Aarhus paper.

In addition to Latour, key figures in ANT include: M. Callon, J. Law, M. Lynch, S. Woolgar, and S.L. Star.

Across the Scales of Time:Artifacts, Activities, and Meanings in Ecosocial Systems

Jay L. LemkeCity University of New York

 

Introduction and Overview

 How do moments add up to lives?

 How do our shared moments together add up to social life as such?

Every human action, all human activity takes place on one or more characteristic timescales. A heartbeat, a breath, a step, a spoken word takes but a moment; a stroll, a conversation extends over many such moments; and an education or a relationship may be a lifetime project. The great cathedrals of Europe were built over many human lifetimes, and the languages and discourse patterns of our communities have developed over still longer times. And yet a conversation consists of many momentary utterances, a relationship may be built of many strolls and conversations together, a building or a social institution is erected by the sum of many individual actions in a community.

How? How do actions or events on one timescale come to add up to more than just a series of isolated happenings? How does a language emerge from many utterances?

How does a community emerge from many people-in-action? On how many different timescales is our social life organized? How does persistent organization on longer timescales constrain the likelihood of events on shorter timescales? How do organizational units and processes on shorter timescales make possible the emergent patternings we recognize at longer timescales?

Why time? Our material world is organized on many scales: space, time, matter, energy, information transfer. In many natural systems there is a strong correlation among these: the quick is also small and light and weak and simple; but in more complex systems, especially those where signs and meaning play a role in behavior and system dynamics, these simple correlations break down.

Classical systems theory is rooted in spatial metaphors and the reductionist project: large systems are to be understood by analyzing them in terms of interactions among smaller component subsystems. Molecules are understood in terms of interactions of atoms, atoms through interactions of smaller particles. Organisms are analyzed by a hierarchy of interactions among units at progressively smaller spatial scales: organ systems, tissues, cells, organelles, and macromolecules. Ecosystems are modeled as interactions among species and abiotic elements; galaxies as interactions among star clusters and individual suns. In all these cases there is a fundamental assumption: units nearer in space are more likely to interact and to interact more strongly (i.e. with greater effect on one another). This assumption imposes a ‘spherical’ topology on the system: relative to any center, items at the same distance scale (i.e. in the same spherical shell) are equally likely to be interaction partners, with the closer ones interacting more and the further ones less.

In many complex systems, however, this assumption fails. Two distant points along the same stream may interact more than two nearer points not linked by the stream. Two distant cells may communicate chemically via the bloodstream; two distant neurons may interact more than closer ones not in the same neural network of pathways, sensitized to the same neurotransmitters or neuromodulators. In a pond or an ocean, two species in the same layer of water, at the same depth may be more likely to interact over wide (horizontal) distances than they are to encounter a species nearer in vertical distance, but separated ecologically by depth-dependent differences of light, temperature, salinity, or pressure. Species roam far in the rainforest canopy without ever venturing a few meters down. In our human ecosocial systems, which are just a specialized kind of ecosystem, people who are linked by the same river, the same railroad, the same phone network, the same chat room on the internet may interact far more than they do with spatially nearer neighbors who are off these social transport and communication networks. In a modern city, spatial proximity may have little relevance to probability or intensity of interaction.

In addition to the ‘spherical’ topology for strength of interaction, there are clearly at least these two others: the laminar topology (horizontal layers) and the network topology (lines of connectivity). All three of these principles are at work in the spatial organization of human ecosocial systems, but I generally agree with Bruno Latour that ‘sociotechnical networks’ are critically important for answering the questions with which I began this article (cf. Latour 1994, 1996; Lemke 1995, 1997). Many people interpret Latour’s arguments, or the somewhat similar arguments proposed by ethnomethodologists (e.g. Garfinkel & Sacks 1970, Schegloff 19..), as leading us to a

‘flat’ view of social systems: that there are only local interactions of people and things, and all the rest (families, institutions, languages, social communities, class conflicts) are contingent and epiphenomenal, essentially unreal figments of our overly fervid sociological imagination. The ‘flat’ view sees only the human scale, indeed only the scale of the moment and the event, privileging that scale in relation to all others. It does not ask how and why events widely separated in time and space seem to re-enact the same patterns; it does not recognize that there are emergent phenomena unique to every level of organization in a complex dynamical system: recurring and typical patterns of interaction that cannot be explained or predicted from analysis of the interacting units.

Neither of these essentially ‘spatial’ views of complex ecosocial systems is satisfactory. The spatial hierarchy model ignores the important role of network toplogies of interaction. The ‘flat’ interpretation of network models cannot account for regularities at higher-scale levels. The ‘spatial’ view is incomplete, and indeed is not, I believe, the fundamental view needed to understand complex systems, especially human ecosocial systems. I want to argue here for the usefulness of an alternative, more dynamical view. 

In dynamical theories of complex systems, the fundamental unit of analysis is a process. It is in relation to the process that its participants are defined, as filling roles in that process. Things, or organisms, or persons, or institutions, as usually defined, are not dynamical notions: they are ordinarily defined in terms of their stable and persistent, or invariant, properties. They are not about dynamics; not about change and doing, but about being what they are. Every process, or action, or social practice, or activity, occurs on some timescale (in complex cases more than one timescale). In a dynamical theory, an ecosocial system is a system of interdependent processes; an ecosocial or sociotechnical network is described by saying what’s going on, what’s participating and how, and how one going-on is interdependent with another. Each scale of organization in an ecosocial system is an integration of faster, more local processes (activities, practices, doings, happenings) into longer timescale, more global or extended networks. It is relative timescale that determines the probability and intensity of interdependence (according to what I will call below the adiabatic principle), and it is the circulation through the network of semiotic artifacts (books, buildings, bodies) that enables coordination between processes at radically different timescales. 

In this view the two fundamental questions for analyzing the dynamics of ecosocial systems – and human activities within them – are: What processes, what kinds of change or doing, are characteristic of each relevant timescale of organization of the system/network? and How are processes integrated across different timescales? In the sections that follow I want to develop in more detail the implications of multiple-timescale analysis for the study of meaningful human activity and to raise a host of research questions generated by this perspective. My principal example will be schooling in relation to identity development and cultural continuity. I will move from a brief consideration of the basic dynamics of complex systems in general, to the case of ecosystems in which meanings matter, and finally to the conclusion that ‘it takes a village’ to study a village.

Activity in Time 

Imagine a school classroom at work. What’s happening? What are the characteristic timescales of the processes and events that make it ‘a classroom’ for us? Almost

certainly people are talking, and their actions, whether producing an utterance or writing at the chalkboard or handing someone a scissors, cannot be understood, either as a selection from the available repertory of human actions or in terms of timing and sequence (what next? what, when?), apart from the meanings being made in talk and other forms of action (for detailed descriptions and analyses of the particular classroom scenes I have in mind as I write, see Lemke 1995, 1996, in press; Kamen et al. 1997; Roth 1998). 

What are the characteristic timescales of the actions, processes, and events we observe? Our immediate human interactional timescale ranges from the glance and the word, said or done in a second or less, to the complex sentence spoken or heard, the complex action performed over a few tens of seconds. Evolution has tuned us to this narrow range of timescales; our survival has depended for hundreds of millennia on noticing brief events (a glimpse of predator through the brush) and sustaining short-term cooperative action. But in the classroom, if we watch and listen long enough, we begin to find repeating patterns (e.g. Mehan 1979, Lemke 1990): individual utterances of certain semantic types (questions, answers, evaluations of answers) predictably follow one another to constitute an exchange. There are identifiable types of exchanges. These recur, recognizably for us and for the participants, not just for a while, or among the same participants, but on different days, in different situations, and even in different classrooms in different schools. They constitute a cultural pattern or social semiotic formation (cf. Lemke 1995). Exchanges also enter into patterns on a still longer timescale; the ebb and flow of talk, the shifts of topic and activity, divide the lesson into episodes. Some types of episodes also recur. Episodes get integrated somehow into lessons, and there are also lesson types, and even sequences of lesson types (cf. Christie 19.. on ‘curriculum genres’) that recur across wide geographical areas and which may take days or weeks to complete. At these longer timescales there are curriculum units and months- to years-long integrated curricula. 

What about shorter timescales? Even an utterance of a single word consists of recognizable and repeated sound patterns, the articulation of the distinctive phonemes of a language, recognizable and repeatable, though perhaps at the edge of our normal awareness. A typical English vowel or short syllable takes about one-tenth of a second to articulate. These articulations represent coordinations of fine muscle control by neuron impulses acting at the few to tens of milliseconds scale, and they in turn sum over neuron membrane depolarizations and ion and neurotransmitter flows that occur on the millisecond timescale. Below that are still faster biochemical reactions, but the scale of the fastest human actions is basically set by the millisecond scale of neuron processes. No coordinated human action, not even autonomic reflexes, can happen faster. 

Table 1 shows the approximate timescales for each order of magnitude above and below the 1 second focal scale of human action. At each timescale we can recognize characteristic processes and social practices. For adjacent timescales it is also quite clear that the processes at the next lower timescale make possible the repeatable patternings of the next longer scale, in accord with the reductionist model of systems hierarchies. What is equally important, however, is that there is always also a higher level process already in place, already running on its own longer timescale, and this sets the context which constrains what is likely and what is socially appropriate at the next scale below. A student’s answer to a teacher’s question is also meaningful for the participants as part

of an exchange, not just as an utterance in its own right, and is judged as appropriate or not to the on-going exchange, and to the episode, the lesson, the unit, the curriculum … and many higher-level contexts. These contexts however are not static, they are themselves processes unfolding in time. Very slow processes function like constant, static backgrounds on the timescale of much faster processes.   

TABLE: Timescales for Human Activity

 

Chemical synthesis 10-5  

Neurotransmitter synthesis

Membrane process 10-4   Ligand binding

Neural firings 10-3   Neuron process

Neuronal patterns 10-2   Multi-neuron process

Vocal articulation 10-1   Edge of awarenessUtterance 1-10 seconds   Word, holophrase, short

monologue; in contextExchange 2- 102 Seconds to minutes Dialogue; interpersonal

relations ; developing situation Episode 103 o(15 minutes) Thematic, functional unit;

speech genre, educativeLesson 103 – 104 Hour Curriculum genre Lesson sequence 104 o(2.75 hours) Macro curriculum genre School day 105 Day [“seamless day”]Unit 106 11.5 days Thematic, functional unitUnit sequence     [rare]Semester/ Yr Curriculum

107 4 Months Organizational level; unit in next scale

Multi-year Curriculum

108 o(3.2 years) Organizational level; limit of institutional planning

Lifespan Educational Development

109 o(32 years) Biographical timescale;Identity change

Educational system change

1010 o(320 years) Historical timescale; New institutions

 Note: o(time) means times on the order of magnitude of ...

Worldsystem change 1011 3200 years New cultures, languages; limit of historical records

Ecosystem, climate change

1012 32,000 years  

  1013 320,000 years Last ice age Evolutionary change 1014 3.2 million years Scale of human evolution   1015 32 million years Dinosaurs   1016 317 million years Pangaea

Planetary change 1017 3.2 billion years Origin of life, of planet Universal change 1018 32 billion years Cosmological processes

  

It is useful to analyze scale hierarchies in groups of three levels at once (cf. Salthe 1985, 1993; Lemke 1995, in press-b). Call the middle level of any such group the focal level, or level N, the focus of our interest for now. Dynamical systems theory basically says that processes (and participants in those processes) can only interact directly and exchange significant amounts of energy or information if they are on the same scale. Technically this is only exactly true if we mean on the same time scale (see discussion of the adiabatic principle below). What is possible on the focal scale, the kinds of interactions that can happen, depends on the kinds of processes and participants at the level immediately below, level N-1. Processes at level N-1 are constitutive of processes at level N; they provide the affordances for activity at level N. But level N is never the top level (certainly for human social processes); interactions on the focal level are not free to range over all the possibilities afforded them: they are also constrained by being themselves part of longer timescale processes at level N+1. The longer-scale processes determine what is probable at the focal level. There are always many ways in which the interactions at level N can satisfy the constraints of level N+1, but the probability of each path depends in part on whether it is consistent with the emerging patterns at level N+1. 

It helps in understanding these inter-level relationships to think of their history and origin. As interactions at some timescale become linked, or coupled, and thus more interdependent, as they do in complex systems of the kind we are interested in here, there are fewer and fewer possible self-consistent patterns (cf. Kaufman 1993). In the origin of life, for example, there were originally a lot of protein or RNA-like molecules mutually catalyzing one another’s synthesis. As more new molecules were produced that could function as potential catalysts for still more chemical reactions, eventually a condition was reached where a few sequences of reactions formed a self-sustaining cycle that then rapidly outpaced other reactions and entrained most of the available nutrients into its own on-going patterns. The new patterns were emergent, essentially allowed by the previous chemistry but not required by it, and so unpredictable. The new cycles take more time, complete on a longer timescale than the individual constituent reactions. They form a new level of organization. From now on, any fluctuations in the concentrations of chemicals in the pool are buffered by the existence of the new cycles, and information which takes this form only reaches some still longer-scale (N+1) process after being filtered or buffered by the new cycles. Once in place the new cycles also alter the probability of reactions occurring on the level below them, providing downwards constraints. And of course the new level now becomes itself a potential unit of organization for something (like us) at a still higher level to be built out of. 

All new levels of organization emerge as intermediates between pre-existing levels, and profoundly change the relations among the formerly adjacent levels as well as making possible still newer emergent forms (see discussion in Lemke, in press). 

Are there emergent processes and patterns in classrooms? I think every teacher and student knows that there are. There are new routines that emerge, new social groupings and the typical interactions that sustain them, class in-jokes, informal rituals, typical

sayings and phrasings, favorite word usages with special meanings, etc. These in turn can become the raw material for more complex new patterns unique to the classroom, and they certainly constrain the probabilities of actions and utterances which would invoke these special meanings or contribute positively or negatively to social relationships. A classroom, and indeed every human community, is an individual at its own scale of organization. It has a unique historical trajectory, a unique development through time. But like every such individual on every scale, it is also in some respects typical of its kind. That typicality reflects its participation in still larger-scale, longer term, more slowly changing processes, which shape not only its development but also that of others of its type. 

A classroom community can be taken as a whole on its longest timescale of activity, typically a few months to a year. We can ask how it develops, as an individual and as a typical instance of its kind. Subject to what constraints from which still larger-scale processes? Made possible by which characteristics of which shorter timescale processes? We can look at its component processes and constitutive units as well, each on their own timescales. And here things begin to get rather complex, because we can easily see that, for example, a social grouping may form that lasts longer than the classroom community. Is it a unit within the classroom community, or a unit at a higher scale? What about a textbook? It is surely the product of a larger-scale, longer-term process (of writing, editing, publishing, etc.), and so a participant in processes on those scales, but it also seems to be a small-scale participant in short-term events within the classroom (e.g. reading aloud a homework problem at the end of a chapter). To understand such phenomena we need to understand the two different principles that seem to govern relationships across scales: the adiabatic principle, and the principle of semiotically mediated heterochrony.

The Adiabatic Principle

There is a fundamental principle in physics, found in slightly different forms in mechanics, thermodynamics, and quantum theory, and generally known as the adiabatic principle. Its basic use in physics is to simplify complex analyses by justifying the neglect of certain possible (but hard to calculate) interactions as being almost certainly too small to make a noticeable difference in the final answer (the adiabatic approximation). ‘Adiabatic’ basically means ‘it doesn’t get through’ referring to energy, fields, or information. In its most basic form it is a statement about energy transfer, and it says that it takes time for energy to be transferred from one system to another; therefore the faster something happens, the less energy is transferred. This means, in effect, that a very fast and a relatively much slower process cannot efficiently communicate with one another, cannot transfer energy. This is the basic warrant for the buffering or filtering effect between non-adjacent levels in the timescale hierarchy, and therefore for the usefulness of defining timescales as being distinct from one another in the first place. 

A process which produces change only very slowly seems to us not to be a process at all, but a constant fact of life. Very slow changes do not produce ‘differences that make a difference’ (Bateson 1972) to us; they do not matter to human life. Weather change processes make a big difference to us, but climate change processes are so slow as to be irrelevant (normally, but that may be changing!). The continents are moving, the Earth’s magnetic poles are shifting, the equinoxes are precessing, the rotation of the earth is

slowing, the energy output of the sun is changing -- but not fast enough to matter to our sense of geography or day and night. 

Or consider very fast processes, much faster than those at our nominal one-second focal level. If you run fast enough across the hot beachsand your feet get less burned because less total energy is transferred to you in the shorter time (for hot coals you may need additional help.) The extreme case was graphically illustrated in a recent film of H.G. Wells’ classic The Time Machine, in which the protagonist survives a nuclear blast in London by accelerating through time at the maximum rate, thus spending too little time in the actual moments of blast energy for very much of it to transfer to him and the machine. Closer to home, fast molecular and atomic processes within the human body do not play a role in our much slower biochemistry, nor can we decipher speech presented to us more rapidly than the maximum rate at which our neurons can respond and process the signals. Moreover, and this goes beyond and adds to the separability of timescales guaranteed by the adiabatic principle, we are buffered from fast, small-scale events, like ionization of individual atoms in our bodies or even errors in gene transcription, by longer term regulatory and self-correcting processes typical of the intermediate scales of autopoietic or self-organizing systems. 

Of course our small degree of autonomy from the environment, within and without, at smaller scales and larger ones, has its distinct limits. One molecular error in one cell can sometimes lead to a cancer that kills the organism. Someday we may cross a threshold in long-term climate change processes and find sudden droughts and famines on a very human timescale. The adiabatic principle has exceptions, and one of these is fundamental to human social organization.

Heterochrony and Semiotic Mediation

According to the adiabatic principle, events in the remote past or processes with much longer characteristic timescales should have little impact on normal human activity. There are exceptions, of course. A fast, microscopic event can trigger a cascade, an amplifying avalanche of consequences that grows to a much larger, longer-term scale. A little more or less neurotransmitter or ion concentration in one small part of a neuron, a single ligand molecule binding to a membrane in a millisecond can trigger a action potential spike that propagates itself down a long axon and triggers other neurons to fire, stimulating a large section of the cortex. No doubt there are many intermediate levels of feedback loops designed to buffer accidental events of this kind from disrupting normal brain patterns, but the basic phenomenon is fundamental to the human nervous system. A single brief event at some time in history may have consequences that ramify down the ages and affect us today on a short timescale. Such phenomena as dynamical thresholds and bifurcations are a normal part of our current picture of complex dynamical systems. Normally their effects are filtered out by the self-organizing meta-stability of various intermediate scale-level processes. 

But there is another kind of exception to the relative insulation between non-adjacent timescales. Rather than a short timescale event having long-term consequences, as above, we can have the case of heterochrony, where a long timescale process produces an effect in a much shorter timescale activity. This is a very common phenomenon in human social activity. I believe it is the basis for human social organization across timescales. 

Before illustrating this point in the significant case of the classroom, let’s enjoy a more colorful initial example. In feudal Japan, members of the samurai warrior class had the right, even the duty, of avenging slights to their and their clan’s honor by beheading, on the spot, an offending commoner. Ready-to-hand in such a situation is the storied and sacred family sword, passed down father to son for centuries, lovingly polished and razor sharp. But the samurai reaches instead for a much less ready to hand, common and ordinary battle sword, not nearly so well-balanced or well-kept, to decapitate the offender. A years-long historical process of cumulating meaning and value envelops the heirloom sword, but this long-term process intersects with and determines action in a very short-term event. The material object itself, the sword, functions in these processes, both long-term and short-term, not simply through its material affordances – the heirloom sword will do just as well or better to cut off a head and is already in hand – but also through the meanings and value it bears. The samurai acts in the situation, not just in relation to present events and material relations, but also in relation to his interpretation of the appropriateness of using the sword and his own education in the traditions of his family and culture – a process on a timescale intermediate between that of the sword’s history and the present event. 

Everywhere in human culture we find this type of heterochrony: longer-term processes and shorter-term events linked by a material object that functions in both cases semiotically as well as materially. The material characteristics of the object also function as signs for an interpreting system of meanings that belong to processes on a very different timescale than that of the event in which the interpreting process is taking place (Peirce 1998; Lemke 1995, in press-b). Leigh Star (e.g. Star & Griesemer 1989) has identified such phenomena for sociotechnical networks as ones in which ‘boundary objects’ circulate through the network, playing different roles in different situations. Typical in these cases are records (e.g. census forms, zoology fieldnotes, ships’ logs) that are created in many short-term events, but then collated in some ‘center of calculation’ (Latour 1987) to create a summary table or a map (which in turn circulates still further in the network), linking these times and places and events both as a material object and as a sign or text. Considered as a whole, the circulation in the network, the completion of a functional cycle of activities (collecting data, summarizing and publishing data) constitutes a longer timescale process, and one that takes place within a more extensive network than does each constituent event. 

In the classroom, we will find student notebooks and class textbooks, but also many other meaning-inscribed material objects that afford heterochrony. In two analyses of science classrooms in recent years I have been impressed by the role of these material-semiotic artifacts. In one (Lemke 1995/97, 1996; see also Roth 1998, Kamen et al. 1997), students are designing and building, in part improvising, towers and bridges built of soda straws, cut with scissors and welded with a glue-gun, while they talk and their activity spawns emergent practices and goals on several timescales. In the other (Lemke, in press-a), a student in his chemistry and physics classes juggles multiple meaning systems: speech, writing, diagrams, graphs, gestures, mime, numbers and algebraic forms; and their associated artifacts: textbook, notebook, chalkboard, overhead projections, talking and moving bodies, a calculator. 

In this lesson, when a teacher asks a question, several students begin looking through their notebooks. The notes they look at now were written days or weeks ago. The answers they give are influenced in part by what they read and how they interpret it in

relation to the question just asked. The notebook, as a material object with semiotic affordances, as a thing that can also be a sign, materially links two events across time and space and so participates in a process on a much longer timescale than either the event of writing or the event of reading that particular note. And in this case so also does the student (see below). At another juncture, the teacher reads aloud from the textbook, writes on the board, and asks a question that would not have been written or asked as it was without the influence of the textbook’s words. Those words, the generic discourse pattern or discourse formation inscribed in the material object of the textbook has an even longer (cultural) history than does this particular material book, or the unique wording of its text. Not only the processes and activities that produced the textbook, but the processes and activities which produced that standard discourse pattern about chemical reactions and circulated it long ago to the textbook’s authors and editors, are now intersecting through the mediation of the book as a material-semiotic object with the much shorter-term events in the classroom episode. 

But it is not just obviously textual records of past events that can function semiotically to mediate heterochrony and the integration of social activities over very different timescales. What about the architecture of the room? The layout of the seats? The size, weight, shape, color, and heft of the scissors? We know that these also influence events in the classroom on smaller timescales. To the extent that they are simply material constraints, or provide material affordances, they represent a slowly changing constant in accordance with the adiabatic principle. But as ‘texts,’ as signs of the values and habits of a culture, as indexical signs of the work of those who built and designed, they also afford information about the culture of schooling. And there is more. 

What about the student? … who is also a material object, a body on and in which can be inscribed – as clothing, tattoos, neuromuscular habit patterns, verbal memories –meaningful signs. Surely this body is also a material-semiotic artifact, a participant in and a product of longer-term social processes, practices, and activities, and one which circulates as a ‘boundary object’ linking one event and time and place with another? Semiotic mediation and heterochrony as the ground of social integration across time and space did not begin with literate texts, or even most likely with specialized record-keeping. While it may remain a just-so story, it is useful I think to imagine a basic dilemma of our remote ancestors. In small bands they roamed a complex environment. Individuals and smaller groups over time traveled different paths, gained knowledge of dangers and opportunities along their personal trajectories. But what was of value to the band as a whole was the accumulating knowledge that came from integrating individual experiences over geographical journeys, but also over time. The ‘oral tradition’ was not just about myth and legend. It was a cumulative knowledge on a timescale and spatial scale that no individual could match; it was maintained and enhanced by integrating short-term processes of discovering, telling, and re-telling across time. And the mediating artifact was the human organism, circulating in its ecosocial networks.

Identities and Trajectories

 

What’s happening to a student in that classroom? What kinds of changes are taking place, and on what timescales?

Schooling is supposed to facilitate certain changes in the behavior patterns of students. Both educators and their critics wonder how lasting these changes are; how far do they carry beyond the walls of the school? On what timescales do we imagine that personal identities change significantly? or habits of critical reasoning? ways of reading, coping with quantitative problems, or interpreting the natural world? attitudes toward potential careers or value choices? 

Do such changes occur in the course of a 40 minute lesson? Even if we imagined that such a change did take place in some ‘breakthough’ moment, would we still count it as a change of the kind I’ve just described if it disappeared the next day? Or the next week? The formation of identity, or even fundamental change in attitudes or habits of reasoning, cannot take place on short timescales. Even if short-term events contribute toward such changes, it is only the fact that they are not soon erased, do not quickly fade – that subsequent events do not reverse the change – which makes it count. It is the longer-term process, including the effects of subsequent events, which determines for us the reality of basic human social development. 

So how could events on the timescale of a conversation or an experiment or reading a story even contribute to identity development? What is the system, or network, within which a notion of ‘identity’ can be defined? It is surely not that of an isolated organism taken at a single moment of time. At the very least, identity must express itself, and that expression in action takes time. An organism as such is not even alive in a single instant of time; its constitutive processes require finite time to occur, and taken as a whole, from its molecular level to its organismic level of integration, it is a multi-timescale dynamical system. From the viewpoint of physics, an organism is alive only across a time interval that necessarily extends somewhat into the future. It can only be observed to be alive across some finite interval of time, and each of its characteristic dynamical properties has some minimum timescale for possible observation. In fact, the self-organization processes which constitute the living organism as being more than the sum of its nonliving parts have a minimum timescale, biologically, of something not much less than a second or two. 

Moreover, beyond that second or two the organism only stays alive by interacting with its environment. It has to release heat constantly and waste chemicals eventually, it has to take in oxygen steadily and nutrients periodically. It would never have come to exist in the first place without having developed in a supportive environment that supplied exactly the conditions that evolution had led its development to depend on. If we think about the organism as such a dynamical system, we need to revise our ideas of it in at least two fundamental ways. First, it is not definable or viable apart from its operating as an element in a larger system, some minimal ecosystem. Secondly, it is not definable at a single instant in time, but only over finite time-intervals, and in fact ultimately only as a trajectory-entity developing and individuating through its interactions with its environment over the whole lifespan course from conception to decay. 

An organism is a biological unit of organization. Its definition says nothing about semiosis, about how it responds to a material environment in ways that depend on its interpretation of things as signs as well as on its direct interactions with their material properties. What is the minimum timescale on which we can observe a dynamical system to be a person? What is the minimum time, and the typical time, for actions that

indicate both that sign-interpretation is playing a role, and that the sign-interpretation is recognizably that of a social persona? A system with a self-conscious identity? 

Just as in the case of our revision of the notion of an organism to both situate it always within a larger-scale system and to look at it over all its relevant timescales (i.e. as a developing trajectory through time), we must do the same thing, for the same reasons, in the case of the definition of a person. A human is not a person apart from social interactions within a community, nor on timescales less than those on which a sense of identity or habits of sign-interpretation develop and are used. Meanings are not made by organisms, but by persons, and they are not made within organisms but within an ecosocial system that minimally includes other persons and the things they make meaning about, and that minimally operates over timescales sufficient for a developing person to come to engage in socially meaningful interactions with others and with the nonhuman surround. 

So what then is a self-conscious personal identity? We might say that it is a semiotic articulation of a person’s evaluative stance toward interactions. It is what we are inclined to believe or doubt, desire or dislike, expect or find surprising, etc. (cf. Lemke 1998 for the semantic dimensions of evaluative stances in language), cast in the romantic folk-language of ‘who we are’, what social types or categories we identify with on the basis of shared values. It is a very complex construct, not usually explicitly articulated; in fact it would be reasonable to say that people do not have stable, unitary identities, but rather than we all learn to interpret certain persistent evaluative stances toward action in these terms and articulate the relevant pieces ad hoc from situation to situation, and not necessarily consistently. We can also construct some ad hoc consistency if we have to, but we don’t usually bother. 

Thus ‘personal identity’ may not be as long-term a phenomenon as we imagine. Like most everything else, it too requires integration across timescales: across who we are in this event and that, at this moment or the other, with this person or another, in one role and situation or another. Nevertheless, evaluative stances or dispositions do develop and change on timescales much longer than that of a classroom lesson. In his extensive research on evaluative dispositions, Bourdieu (1979) finds that they are created by our participation in the typical activities of a local social community over timescales of a decade or two, and change more slowly as we grow older. He finds significant social class differences, tied specifically to education and to economic life-prospects, and comes close to proposing that social class itself be redefined in terms of the differences in ‘habitus’ or dispositions to action produced by different trajectories of socialization. 

Every developing biological system develops partly uniquely (individuation of the trajectory) and partly as a typical member of its kind (type-specific, equifinal trajectory). The latter effect is due not just to common genetic heritage, obviously, but to similar epigenetic circumstances eliciting gene expression. In fact it is easy to imagine the effect of much of that epigenetic information as just an internalization of the landscape we write with our lives. Each of us leaves some imprint on the world, if only in the bodies and memories of those we interact with; and those imprints, as semiotic mediating artifacts, provide informational input to the development of others of our kind. It is the larger-scale social system, obviously, that persists from generation to generation and tends to make one person in a community grow up to act somewhat like others of his or her time and place and class and gender and age. 

The classroom then is no different from anywhere else in our world of social artifacts. Its developmental input is there not only on the walls but in the very fact that there are walls; not just in the words in the textbook, but in the existence and use of textbooks. But it is first and foremost in those respects in which the classroom is exactly like the rest of the social world that it contributes to the formation of identities and habits of action that are formed across the longer timescales we also spend in other places. It is not what is unique about classrooms that contributes to our identity development, but what is the same about them compared to many other sites in our culture. Identities develop over long timescales, during which the trajectory of the developing social person takes him and her from classroom to classroom, from school to schoolyard, to street corner, to home, to shopping mall, to TV worlds. The timescale for sampling all these worlds that is relevant to identity development is the long timescale, one that sees the sameness of patterning across all these venues. The little differences between them are blips in this long, slow process. Of course we also learn those differences, and the appropriate roles for classrooms and other places, but our more general dispositions are necessarily a function of their commonalities. 

The most amazing feature of developmental processes is that each step along a developmental trajectory changes the way the system interacts with its environment at the next step. There are no ‘shortcuts’ in development; you must pass through each step in order to be prepared to take the next one, because at each step you become a dynamically different system. Different dynamical possibilities are open to you. You have also extended your trajectory to a new timescale on which there are emergent phenomena, in you and in your interactions with a larger-scale environment. In biological organisms like ourselves, the developmental pathway is extremely long and complex, and each critical turn could easily go down some other path; the message from our genes is a roadmap for the paths followed by our most successful ancestors. Our ontogeny recapitulates their phylogeny, up to a point. But only up to a point, and less so as developmental pathways come to be guided more by social interaction and culture-specific semiotic information supplied after birth. 

Everyone in that classroom was experiencing a different lesson, was interacting with the teacher and the semiotic artifacts of the room and with each other in ways that depended on their trajectory-up-to-now (and now-in-progress). No matter how much we homogenize classroom groups -- by age, by social class, by gender, by culture, race, or dominant language – for the classroom processes at each timescale there will be considerable differences in affective engagement, in evaluative dispositions, in relevant knowledge and skills, in resources for integrating the events of the moment into patterns that will persist on longer timescales. The very act of homogenizing defeats the goal of long-term results: the world outside the classroom is not homogenous in any of these ways, and every difference between the meaning organization of the classroom and that of the rest of life means that much less long-term and wide-ranging persistence of what happens in the classroom. (For similar arguments about identity issues and the role of real-life apprenticeship in learning, see Lave & Wenger 1991). 

Nonetheless, some contribution toward identity development is taking place all the time, including during classroom lessons. What’s happening? 

Again it is useful to analyze on multiple timescales. On each timescale each student is participating in some ecosocial processes and taking on relevant roles. Students interact

with one another and with the other available semiotic objects in various intersecting activities, and these activities are recognizable and repeatable and usually repeated. In this participation we learn to do differently and to be different. We engage with a person or an artifact in a particular way, typical of that activity, and now the system in which our persona exists and functions changes. Dynamically, we are what we do, and we are now creating ourselves as personae in interaction with new others and artifacts, which means that the current, and perhaps temporary, ‘I’ is the one that exists in the ‘loop’ of efference and afference, of ‘differences that make a difference’ in a kind of complex feedback circuit (in the terms of Bateson’s 1972 cybernetic version). There are longer-term Selves already engaged in on-going longer-term projects and activities, and the shorter-term Selves of current activities, some of which contribute to longer-term projects and some of which may not. As we interact socially at the human event scale, we ‘identify’, if not with the Other as such (cf. Van de Vijver, in press), at least with our agency and participation in each emergent new activity whole, always taking place in a larger-scale system than our former, or more isolated Self. In fact we can even take a reflective perspective in the activity and see our own role in it, that is we can frame a separated ‘me’ from the viewpoint of this new dynamical ‘I’. Reflexivity is itself an instance of heterochrony. 

But all such activities come to an end. What then of the dynamical ‘I’? is there a longer-term residual effect of our participation? Perhaps only a weak one, for now we are asking about quantitative matters of degree. Will we re-engage in the same activity with the same persons or artifacts? How soon? How often? Will we re-constitute some features of the former activity: the same person in a different but similar activity? or the same artifact in a new activity? the same type of activity but with other participants? And in each case how much of an ‘impression’ will be made on the organism and in the larger system that enables the organism to reconstitute its emerging identities by getting these activities going again, or interacting again with the same persons or artifacts, or ones it considers similar for this purpose? How strong will the affective engagement be? How positive will the identification be, evaluatively? And above all, how long will the sequence of activities last across which the same identity features are being reinforced? 

The person we become for a moment with a new stranger for whom we have no strong feelings and whom we never see or remember again may be transient indeed. The person we feel ourselves to be when interacting with someone we feel strongly about, again and again over the course of a lifetime, is an essential part of who we are. The Self I am when I am writing, or teaching, or doing those things that mean something fundamental to me, and that I can do over many years, is basic to my identity. Even the Self I am when I read a particular book, hear a particular kind of music, play or sing or dance to that music, if I feel strongly enough about it, can become basic to my identity. When I teach, or write, or have conversations with colleagues, I am often working to recreate activities and senses of Self that are basic to my identity. I am seeking to keep an identity-constituting process going on a longer timescale and across a wider range of settings and participants. If my identity’s dispositions value aggressiveness, I may seek activities and roles in which aggression is socially acceptable, or I may make use of other activities in ways that support this identity need. It is likely in our society that gender identity needs are very strong in many people, particularly males, and that the conduct of many life-activities, from driving a car to writing a scientific paper, may be entrained in the service of these needs, and so be a very different experience for males and for females, or generally for those constructing different kinds of gender identities

with different degrees of intensity. (What I say here of gender identities will also be true in general of age-, class-, and more broadly of all subculture- and culture- specific identity patterns.) 

It is somewhat ironic that classroom education and the formal curricula that are supposed to create more long-term continuity from lesson to lesson and unit to unit (though not, after the earliest years, from hour to hour in the same day or from year to year in the same subject) are narrowly focused on informational content which is more or less unique to school experience, when the major developmental processes of these years appear to be about the formation of identities that fit large-scale social models for gender-, class-, age-, and culture- specific patterns. Students are mainly going about the business of learning to be six-year-olds or twelve-year-olds, masculine or feminine, gay or heterosexual, middle-class or working-class, Jewish or Catholic, Irish American or Jamaican American, or any of the many dozens of sociotypical identities for which there are identity-kits available in a particular community (cf. Gee 1992). Whatever we offer up in the classroom becomes an opportunity to pursue this longer-term agenda of identity building; our primary affective engagement is with this agenda, with becoming who we want to be, not with learning this or that bit of curriculum, except insofar as it fits our particular agenda or insofar as ‘being a good student’ or ‘not falling for that bullshit’ fits it. Perhaps late in schooling a few of us are also working to form, within these larger identity projects, specific partial identities as ‘future scientists’ or ‘future teachers’. (How many? How intensely? On how long a timescale? How integrated with the more general identifications like gender?) 

Nor of course does this picture change very much ‘after school’ (in either sense). In our paid employment, as in our family and leisure life, we are often still involved in the long-term project of maintaining and enhancing, and perhaps occasionally revising, who we are. We integrate each event into this longer timescale project in many ways: in our body hexis, in our habitual ways of talking, in our retrospective narratives about ourselves, in written diaries, in titles on bookshelves or inscriptions on trophy walls, in photographs of children and friends, in collections of videotapes and knickknacks. 

But it is not easy to study lives over the timescale of decades and lifetimes.

‘It takes a village …’

 

I want to conclude with some observations about the research process itself. 

I began with two key questions: How do moments add up to lives? How do our shared moments together add up to social life as such? I’ve said a fair bit about the first one now, but the focus on identity development, while clearly situating it in larger-scale social activity and process, has left less said about the timescales relevant to the formation and maintenance of community as such. 

Just as we have mainly studied social development over relatively short timescales (generally the first few years or first decade or two of life), and so know far too little about what human projects are sustained over many decades or a lifetime, so also we know a great deal more about short-term social processes: conversation, negotiation,

‘service encounters’, classroom lessons -- events that last on the order of the time you can record on a videotape -- than we do about activities and processes that last days or months or years, much less multiple lifetimes. The classic model of Activity Theory (Leontiev 1978) distinguished three timescales: operations, actions, and activities, rather similar to the levels exemplified in Table 1 by vocal articulations, utterances, and lessons. But can we lump together all the timescales of ‘activities’ that last from minutes to lifetimes? Are the principles of integration across events just the same whether we are taking out the garbage or building a business empire? At some very high level of abstraction perhaps they are, but we should now become more interested in the potential differences. Following on in this intellectual tradition, cultural psychologist Michael Cole (1996) proposes including a much wider range of scales of analysis, from the microgenetic (event scale), meso-genetic (extended activity or project scale), and ontogenetic (developmental-biographical scale), to the historical and evolutionary scales. He has a particular interest in the emergence of sustainable institutions that persist over times longer than the participation of any one individual in them. 

But how do we study ecosocial processes on timescales longer than a few hours? While educational researchers have done extended videotaping of the same classroom over a whole term or year (e.g. Christie 19..), or even followed the same student through all classes for a few days (e.g. Wyatt-Smith & Cumming, in press), we do not follow students home through all 24 hours of their lives for extended periods, nor do we follow even classroom communities over their whole (short) lifespans by looking not just at what happens in the classroom, but what happens among the same participants outside it as well. When the timescale expands, either with its focus on an individual (cf. the social science fantasy depicted in the film The Truman Show (Weir 1998), or on a group or small community, so usually does the spatial scale as well. People move around. How many settings do the members of a community collectively inhabit in the course of a day, much less a year? How many additional, peripheral participants become involved? Even if we had the resources and the persuasive powers to record data on these scales, how could we ever analyze it all? Or even view it all? Imagine the task for detailed records of daily life for even fifteen people over just a year? 

Ethnographic studies of small village societies of a few to several dozen members normally collect data over a yearly seasonal cycle. It is highly selective data and is supplemented by the ethnographer’s direct participation with a subset of the members -- and some response to the inevitable identity shift as the ethnographer begins to becomes another social person in a very different social system, even on a one-year timescale. 

Biographers take the human lifetime as their timescale unit, and otherwise historians and archivists are about the only scholars who concern themselves with social projects or activities that occur on timescales from decades to longer than a single human lifetime.

We can no doubt learn a great deal from each of these fields about the kinds of projects and activities that occupy the longer timescale spaces on our chart, from those that extend through significant portions of the life of an individual, to those which are undertaken by the members of an institution or of a smaller or larger community. What are the longest timescale projects ever sustained by a human community? How should we define the continuity of such projects? What are the means by which integration across timescales is defined in long-range institutional and community projects? 

The original logic of Leontiev’s Activity Theory defined the continuity or unity of extended activity by the maintenance of a goal or object-of activity. One can argue that in many forms of social activity, goals are emergent; they change during the development of the project (cf. Lemke 1996). They are also, in collective activity, not necessarily common or shared among participants; different goals are just successfully enough articulated to permit collective activity to proceed for the most part coherently. When the timescale of a project or activity exceeds a single human lifetime, we clearly should be seeking for alternative principles to define their unity and continuity. Indeed we also have to recognize that many social processes are simply emergent in communities on these timescales: they just happen without anyone or any group intending them. This, too, is characteristic of self-organizing systems on all scales. 

How, finally, do we study society? Or more properly, a whole ecosocial system? And study it moreover from the inside, which is contrary to the externalist tradition of modern science (cf. Matsuno & Salthe 1995) which looks at all systems from a ‘God’s eye’ view. Traditional macrosociology has resorted, after the manner of Latour’s ‘centers of calculation’ to assembling statistical data and recognizing that it does so in a positioned way. The kinds of data we seek to collect are usually the ones that seem important from where we sit within the system. It is highly unlikely that any social system looks the same from the viewpoint of all the component groups or roles within it. Ethnographers long ago recognized that men’s and women’s views of even small ‘homogeneous’ societies are very different. We can say the same in many cases for the views from different age-groups, social classes, or minority cultures. 

There is still a strong individualist bias in our modernist traditions of research. Whether we favor individualist or communitarian politics, agent-centered or collective-process models, we still tend to define our objects of study in such a way that a single researcher could in principle come to understand them. This appears to be a contradiction in the case of ecosocial systems. The longest timescale processes that characterize such systems are almost certainly longer than a human lifetime. We cannot study such a system from more than a few of the many viewpoints within it, and we honestly do not expect all these views to fit consistently together. We need at least a team to conduct such a study, one as diverse or nearly so as the system under study, and along the same dimensions of difference. And we need a self-sustaining institution that will last long enough to observe major historical change in the system. ‘It takes a village’ to study a village. 

Is this possible? There is perhaps some hope insofar as distributed communities of researchers, linked by new communication networks and technologies, may grow to become such ‘villages’ and continue their work over the timescales needed. If we are really optimistic, we might even imagine that before too long all the various viewpoints in our increasingly global society might come to be represented in one way or another in such a research community. But that will mean that the nature of the research project itself will have to change quite radically from how we conceive it today. Our views of social research – its goals, methods, and objects of study – are inevitably still masculinized, still middle-class, still eurocultural, still specific to the interests of a particular age-range within our own social system, in ways we can just barely begin to perceive. And what of the additional viewpoints of the ‘nonhumans’ as Latour calls them – the other participants without which there would be no system? What new roles

should they play in an ecosocial community that is able to understand itself across the scales of time?

 END.

References

Bateson, G. 1972. Steps to an ecology of mind. New York: Ballantine.

Bourdieu, P. 1979. Distinction: A social critique of the judgment of taste. Cambridge, MA: Harvard University Press.

Cole, M. 1996. Cultural psychology. Cambridge, MA: Harvard University Press.

Christie, F. 1997. Curriculum macrogenres as forms of initiation into a culture. In Christie, F. and Martin, J.R. (Eds.), Genre and institutions: Social processes in the workplace and school (pp.134-160). London: Cassell.

Garfinkel, H., and Sacks, H. 1970. On formal structures of practical actions. In J.C. McKinney and E.A. Tiryakian (Eds.), Theoretical sociology: Perspectives and developments (pp. 337-366). New York, NY: Appleton-Century-Crofts

Gee, J. P. 1992. The social mind. New York: Bergin and Garvey.

Kauffman, S. 1993. The origins of order. Oxford: Oxford University Press.

Latour, B. 1987. Science in action. Cambridge, MA: Harvard University Press.

Latour, B. 1994. On technical mediation. Common Knowledge 3(2): 29-64.

Latour, B. 1996. On interobjectivity. Mind, Culture, and Activity 3(4): 228-245.

Lave, J. & Wenger, E. 1991. Situated learning: legitimate peripheral participation. Cambridge: Cambridge University Press.

Lemke, J.L. 1990. Talking science: Language, learning, and values. Norwood, NJ: Ablex Publishing.

Lemke, J.L. 1995. Textual politics: Discourse and social dynamics. London: Taylor & Francis.

Lemke, J.L. 1995/1997. Making towers, making withs. Paper presented at National Association for Research in Science Teaching, San Francisco, April 1995. Arlington VA: ERIC Documents Service (ED 384 513), 1995. Included in M. Kamen et al., Multiple perspectives on the role of language in science education. EJSE: Electronic Journal of Science Education, 2(1): http://unr.edu/homepage/jcannon/ejse/kamen_etal.html

Lemke, J.L. 1996. Emergent agendas in collaborative activity. Paper presented at American Educational Research Association Annual Meeting, San Francisco

(April 1995). Available via Arlington VA: ERIC Documents Service (ED 386 425), 1996.

Lemke, J.L. 1997. Cognition, context, and learning: A social semiotic perspective. in D. Kirshner and A. Whitson, Eds., Situated cognition theory: Social, neurological, and semiotic perspectives. Hillsdale, NJ: Erlbaum. Pp. 37-55. 1997.

Lemke, J.L. 1998. Resources for attitudinal meaning: Evaluative orientations in text semantics. Functions of Language 5(1): 33-56.

Lemke, J.L. In press-a. Multimedia literacy demands of the scientific curriculum. Linguistics and Education 11, 1999.

Lemke, J.L . In press-b. Opening up closures: Semiotics across scales. In J. Chandler and G. van de Vijver (Eds.), Closure: Emergent organizations and their dynamics. New York: Annals of the New York Academy of Sciences.

Leontiev, A.N. 1978. Activity, consciousness, and personality. Englewood Cliffs, N.J. : Prentice-Hall.

Matsuno, K. & Salthe, S. 1995. Global idealism, local materialism. Biology and Philosophy 10: 309-337,

Mehan, H. 1979. Learning lessons: Social organization in the classroom. Cambridge, MA: Harvard University Press.

Peirce, C.S. 1998. The essential Peirce: Selected philosophical writings, vol. 2. Bloomington, IN (USA): Indiana University Press.

Roth, W.-M. (1998 ). Designing communities. Dordrecht, Netherlands: Kluwer Academic Publishing.

Salthe, S.N. 1985. Evolving hierarchical systems. New York: Columbia University Press.

Salthe, S.N. 1993. Development and evolution. Cambridge: MIT Press.

Schegloff, E.A. 1991. Reflections on talk and social structure. In D. Boden and D.H. Zimmerman (Eds.) Talk and social structure. Berkeley, CA: University of California Press.

Star, S.L. & Griesemer, J.R. 1989. Institutional ecology, 'translations' and boundary objects: Amateurs and professionals in Berkeley's Museum of Vertebrate Zoology, 1907-39." Social Studies of Science 19: 387-420.

Van de Vijver, G. In press. Psychic closure: A prerequisite for the recognition of the sign-function? Semiotica.

Weir, P. 1998. The Truman show. Hollywood, CA: Paramount Pictures. [Film]

Wyatt-Smith, C. & Cumming, J. In press. Examining the literacy-curriculum relationship. Linguistics and Education 11 (Special issue: Analyzing the literacy demands of the curriculum).

ANALYSING VERBAL DATA: PRINCIPLES, METHODS, AND PROBLEMS

J.L. LEMKE

Increasingly, the data of science education research are verbal data: transcripts of classroom discourse and small group dialogues, talk-aloud protocols from reasoning and problem-solving tasks, students' written work, textbook passages and test items, curriculum documents. Researchers wish to use data of these kinds to describe patterns of classroom and small-group interaction, development and change in students' use of technical language and concepts, and similarites and differences between school and community cultures, school science and professional science, the mandated curriculum and the delivered curriculum. In a short chapter is it not possible to demonstrate actual state-of-the-art techniques of linguistic discourse analysis. My purpose here will be to formulate the issues and choices of which researchers should be aware in adopting and adapting any method of analysis of verbal data for their own work. Along the way I will cite examples from my own published work and other sources which I personally find useful. Discourse analysis is a very large subject; its principles embody a theory of meaning-making that is nearly co-extensive with a theory of human behaviour and human culture (Lemke 1995a). (For other useful introductions to discourse analysis and classroom discourse study, see Brown & Yule 1983; Cazden 1986, 1988; Coulthard 1977; Coulthard & Montgomery 1981; Edwards & Westgate 1994; Stubbs 1983; Widdowson 1979; Wilkinson 1982)

How Researchers Construct Verbal Data

The language people speak or write becomes research data only when we transpose it from the activity in which it originally functioned to the activity in which we are analysing it. This displacement depends on such processes as task-construction, interviewing, transcription, selection of materials, etc., in which the researcher's efforts shape the data. Because linguistic and cultural meaning, which is what we are ultimately trying to analyse, is always highly context-dependent, researcher-controlled selection, presentation, and recontextualisation of verbal data is a critical determinant of the information content of the data. Data is only analysable to the extent that we have made it a part of our meaning-world, and to that extent it is therefore always also data about us. Selection of discourse samples is not governed by random sampling. Discourse events do not represent a homogeneous population of isolates which can be sampled in the statistical sense. Every discourse event is unique. Discourse events are aggregated by the researcher for particular purposes and by stated criteria. There are as many possible principles of aggregation as there are culturally meaningful dimensions of meaning for the kind of discourse being studied.

The basis for aggregation, ultimately, is covariation: some change in the context or circumstances is associated with a systematic change in discourse features of interest to the study. Normally this cannot be known until the end of the study, so it is wise to

collect a larger and more diverse corpus of verbal data than will ultimately be used to support the analysis. The basis of discourse analysis is comparison. If you are interested in covariation between text features and context features, you should not collect data only for the cases of interest, but also for cases you believe will stand in contrast with them. If, for example, you are interested in phenonema specific to women, to third-graders, to small-group discussions in lab settings, or to a particular curriculum topic, you should also collect potential comparison or reference data, in small amounts, for other genders, grades, settings, or topics.

Discourse analysis is also contextual. If you are interested in the language of any particular kind of event or text, you should also collect "around" it its probably relevant intertexts (see below). If you are studying how students write up their lab work, in addition to the texts they write, you will also need data on how the same topics have been discussed in whole class sessions, what the textbook says on the topic, any relevant written handouts, and perhaps also interviews with the teacher and the students. All analysis is reductive. Information from the original data is discarded in the process of foregrounding the features of interest. Wise researchers preserve the original data in a form that can be re-analysed or consulted again from a different viewpoint, posing different questions.

Spoken language is never analysed directly. It is not even often analysed directly from audio or video recordings, but from written transcriptions. The process of transcription creates a new text whose relations to the original data are problematic. What is preserved? What is lost? What is changed? Just the change of medium from speech to writing alters our expectations and perceptions of language. What sounds perfectly sensible and coherent may look in transcription (any transcription) confused and disorganised. What passes by in speech so quickly as not to be noticed, or is replaced by the listener's expectations of what should have been said, is frozen and magnified in transcription. Normal spoken language is full of hesitations, repetitions, false starts, re-starts, changes of grammatical construction in mid-utterance, non-standard forms, compressions and elisions, etc.

The tendency in transcription is to "clean it up", dismissing most of these features as irrelevant. Very often some of them turn out not to be irrelevant at all. I recommend transcribing large portions of the corpus at the "lexical" level (preserving the sequence of whole, meaningful words and meaningful non-lexical vocalisations) for survey purposes, and smaller portions at more detailed levels for more intensive analysis. The simplest transcriptions attempt to preserve information at the level of the word, but language only occasionally constructs meaning with single words. What matters is how the words are tied together, and that often includes intonation contours. Whether two phrases represent self-paraphrase or contrasting meanings often can only be determined from intonation. Transcription at the level of the word also erases information about emphasis, value-orientation, degree of certainty or doubt, attitude of surprise or expectability, irony, humor, emotional force, speaker identity, and speaker dialect or language background. Many of these features are often redundantly coded in the words as well, but some may not be. In addition, information about the timing of speech (length of pauses, simultaneous speech, sudden breaking-off of fluency, overlaps, etc.) is often important.

Written texts carry considerable visual information: handwriting forms, page layout, typography, accompanying drawings and illustrations, etc. This information, which can be very important for interpreting the meaning of verbal text, should not be lost to the analysis. Videotapes obviously contain a wealth of relevant visual information on gaze direction, facial expression, pointing and other gestures, contextual artifacts referred to in the verbal text, positional grouping, relative distances and directions, etc. Along with fieldnotes, they help us to reconstruct the social situation or cultural activity type within which some meanings of the verbal language are very much more likely than others. For useful discussions of transcription, see (Ochs 1979; Sacks et al. 1974). For the role of intonation, see (Halliday 1967, Brazil et al. 1981). On visual information in text, see (Bertin 1983, Kress & van Leeuwen 1990; Lemke in press-a; Tufte 1983).

The Contexts of Verbal Data

Language is always used as part of a complex cultural activity. Verbal data make sense only in relation to this activity context and to other social events and texts with which we normally connect them, their intertexts. Meaning is not made with language alone. In speech it is accompanied by gestural, postural, proxemic, situational and paralinguistic information; in writing by choices in the visual coding of words and other graphical information. The meaning of any text or discourse event always depends on how we connect it to some (and not other) texts and events (on general intertextuality see Lemke 1985, 1988a, 1993). What the teacher is saying now makes sense in part in relation to what she said ten minutes ago or yesterday, what we read in the book, the question you missed on the last quiz, etc. It also makes sense differently depending on whether she is reviewing or introducing new material, whether it is addressed to one student or to the whole class, whether it relates to a diagram on the board or not. What a student says may make meaning in relation to the past history of his dialogue with this teacher, the group dynamics of the class, his boredom with the topic, his personal relations with other students.

There are many schemes for systematising the probably relevant contextual factors of a text or discourse event (see for example Erickson & Shultz 1981; Hymes 1972). They all include: the participants and their social and physical relationships, material objects and semiotic representations in the immediate physical environment, the cultural definition of the activity type or situation and its roles and expectations, and the channel or medium of communication. More important than such lists are (1) the principle that the discourse itself can create a context, make a part of the environment newly relevant, or even change its meaning; and (2) that the context is itself a kind of text, it must be "read" from the viewpoint of the verbal discourse. Verbal data, including particularly written or printed texts, always makes sense in relation to (1) a context of production, the circumstances in which it was written or spoken, and (2) a context of use, those in which it is read or heard. For written texts these two can be very different (see Lemke 1989a). Texts and discourse data index or point to relevant contexts in a variety of ways (see Silverstein 1976; Wortham 1992, 1994). The simplest is through deictic forms such as this, that, the other, over there, now, as we saw before, mine etc. These forms indicate to the listener that meaning must be made jointly with the textual and the relevant contextual information. In addition to the context of situation, there is also more generally the context of culture (see Firth 1957; Halliday & Hasan 1989; Hasan 1985; Malinowski 1923, 1935) that is indexed by a text. Much of this is a presupposition of familiarity with other texts, cultural norms, genre conventions (see

below), etc. in a particular community. Nonverbal signs which co-occur with spoken language, especially "body language" signs form, with speech, a single integrated meaning-making and interpersonal communication system. Very little is really known yet on how the different channels of this system modulate each other's meaning effects, but see Kendon 1990; Lemke 1987; Scheflen 1975.

The Dimensions of Verbal Meaning

Language in use always creates three interdependent kinds of social and cultural meaning. It constructs social relationships among participants and points-of-view; it creates verbal presentations of events, activities, and relationships other than itself; it construes relations of parts to wholes within its own text and between itself and its contexts.

Presentational meaning is the most familiar and most studied. This aspect of meaning is often referred to as representational, propositional, ideational, experiential or thematic content. This is the function of language for presenting states-of-affairs, for saying what is going on. It presents processes, activities, and relationships; the participants in these processes, and attendant circumstances of time, place, manner, means, etc. It defines entitites, classifies them, ascribes attributes to them, counts them. In relation to these semantic functions, its grammar has been usefully described by Halliday (e.g. 1976, 1985; see also Martin 1992). My own work on thematic patterns or formations (Lemke 1983, 1988a, 1990a, 1995b) applies Halliday's analysis to textual and intertextual patterns in discourse (see below).

Orientational meaning may be even more fundamental developmentally. This aspect of meaning, also called interpersonal or attitudinal, constructs our social, evaluative, and affective stance towards the thematic content of our discourse, towards real and potential addressees and interlocutors, and toward alternative viewpoints. It includes the language of formality/intimacy, status and power relationships, role relationships; speech acts such as promising/threatening, joking, insulting, pleading, requesting/demanding, offering, etc.; evaluative stances toward the warrantability, normality, normativity, desirability, seriousness, etc. of thematic content; construction of affective states; and construction of alliance, opposition, etc. between one theory or viewpoint about a matter and others available in the community. Useful sources on these aspects of orientational meaning include: Austin 1962; Bakhtin 1935, 1953; Halliday 1978, Hasan 1986, Hasan & Cloran 1990; Martin 1992; Lemke 1988a, 1989b, 1990b, 1992; Poynton 1989; Thibault 1991.

Organisational meaning is not always perceived in our culture as meaning, but analysis shows that it is an integral member of the team, functioning together with, and indeed enabling, the other two. Organisational meaning includes the ways in which language creates wholes and parts, how it tells us which words go with which other ones, which phrases and sentences with which others and how, and generally how a coherent text distinguishes itself from a random sequence of sentences, phrases, or words. Organisational meaning in language is generally created through simultaneous use of two complementary principles: (1) constituency structure, in which a larger meaning unit is directly made up of contiguous smaller units, and (2) cohesive structure, or "texture", in which chains of semantic relationships unite units which may be scattered through the text. Constituency structures may be interrupted and resume, and are at least

in principle "completable". Cohesion chains, which have neither of these properties, are built on a variety of chain-membership principles, all of which specify a particular kind of relation of meaning among the items (e.g. synonyms, members of a common class, contrast, agent-action, action-means, attribute-item, etc.). Constituency structures (genres, genre stages, rhetorical formations, adjacency structures, clause-complexes, clauses, phrases, groups, etc.) create local meaning relationships among items which also generally belong to cohesion chains, and they provide one means for creating new bases for cohesive relations. Real texts, especially extended complex discourses, often change genre types or other constituency strategies many times, creating sub-units within a text. Cohesive relationships provide a principal means of creating semantic continuity across these segmental boundaries within a text. Note that some forms of meaning depend about equally on two of these three semantic functions, so that, for example, logical relationships (because, if ... then) normally function both presentationally and organisationally. For useful discussions of organisational meaning, see: Halliday 1978; Halliday & Hasan 1976, 1989; Hasan 1984; Lemke 1988b, 1995b; Martin 1992; Matthiessen 1992).

Semantic Content Analysis

How can we characterise what a text says about its topics, or even what its topics are, better or more concisely than the text does itself? This is possible only to the extent that the text repeats the same basic semantic patterns, makes the same basic kinds of connections among the same basic processes and entitities again and again. It happens, in our culture, and probably in most, that not only do we repeat these thematic patterns, or formations, again and again in each text, merely embroidering on the details, we also do so from one text or discourse event to another. This is especially true in the sciences and other academic subjects where there are accepted, canonical ways of talking about topics. Most textbooks will tell you pretty much the same thing about atoms, or alternating current, or Mendelian inheritance. We expect that, however they present it, what teachers say about these topics will contain this same information, and that when students reason, talk, write, or take tests, that their discourse will fit these patterns, too (at least eventually).

The common techniques of concept mapping are based on our ability to consciously abstract the essential meaning relations among key terms in scientific discourse. Discourse analysis, however, can produce the same patterns, and be more semantically explicit about their content, from free-form classroom or small-group talk, or from written materials of any kind. This means that these direct uses of scientific concepts can be directly sampled, assessed, and compared. The basic technique for doing this is described in Lemke (1990a) and its linguistic basis and extensions are discussed more fully in Lemke (1983, 1988a, 1995b). Other forms of modern semantic content analysis are statistical, corpus-based, and collocational (see e.g. Benson & Greaves 1992; Halliday & James 1993; Sinclair 1991). Given the present limitations of computer analysis of natural language texts, these analyses are based on forms rather than meanings. They can tell you the frequency distributions, and more importantly the joint distributions for pairs (or n-tuples) of words or fixed phrases, in a text. They cannot tell whether a given word is used with the relevant meaning you are interested in in any particular instance. Thematic analysis, correspondingly, must be done by hand, but it enables you to see that the same concept or relationship may be expressed by many different verbal forms and grammatical constructions, and to exclude cases where the

form is right but the meaning in context is not. To do thematic analysis properly, you need to be familiar with both the subject matter content of the discourse or text, and with the semantics of at least basic lexical and grammatical relations at the level of Halliday (1985) and Hasan (1984).

Rhetorical Interaction Analysis

All language in use, whether spoken or written, is explicitly or implicity dialogical; that is, it is addressed to someone, and addresses them and its own thematic content, from some point-of-view. It does rhetorical and social work, producing role-relationships between author-speaker and reader-hearer with degrees of formality and intimacy, authority and power, discourse rights and obligations. It creates a world of value orientations, defining what is taken to be true or likely, good or desirable, important or obligatory.

Some useful questions to guide rhetorical analysis include: What are these people trying to accomplish here? What are they doing to or for one another? How is the talk ratifying or changing their relationships? How is it moving the activity along? How is it telling me what the speaker/writer's viewpoint is? What is it assuming about my viewpoint? other viewpoints? How does it situate itself in relation to these other viewpoints? What is its stance toward its own thematic content, regarding its truth or probability, its desirability, its frequency or usuality, its importance, its surprisingness, its seriousness, its naturalness or necessity? Rhetorical analysis needs to be done at each organisational level of the text. What is the function of the choice of genre as a whole (see below)? of each stage in the unfolding of the genre? of the local rhetorical formation and each move within it? of the sequencing of formations and topics? of various interruptions, digressions, and the timing of returns? Of grammatical contructions? Of word choices? Of pauses, intonations, marked pronunications?

Given a particular thematic content, there are an endless variety of grammatical (including non-standard) ways to word it. Each variation fits the need of some rhetorical situation, or helps us to construct one. While genres or common rhetorical patterns provide a definite set of expectations, they also allow or encourage considerable strategic and tactical manoevering (see examples in Lemke 1990a, chapters 1 and 3). It is not always possible to say what a particular choice or move means, but you can say what it might mean. And frequently it means more than one thing, plays a role in more than one action or strategy, even in unconscious ones. Those features of a rhetorical analysis which rely, as thematic analysis does, on patterns that are commonly found in many texts, tend to be agreed on by different analysts. But rhetorical analysis must deal with situations unique to the text at hand much more often, and these are more ambiguous and subject to different interpretations. In these cases multiple forms of evidence need to be used to support interpretations: word choice, intonation, grammatical choice, contextual information about the situation or activity. Even the participants in a discourse may disagree about the rhetorical meanings of particular features, or change their minds in retrospect or with additional information. The "intention" of the speaker, revealed in a retrospective interview, is just one more piece of data; it does not settle the question of what a feature meant for any participant at the time. Evidence of how participants followed-up on the appearance of the feature may be more persuasive.

Discourse forms do not, in and of themselves, "have" meanings; rather they have a range of potential meanings. Words, phrases, sentences are tools that we deploy in complex contexts to make more specific meanings, to narrow the potential range of possible meanings down to those reasonably or typically consistent with the rest of the context. Even in context, at a moment, an utterance or phrase may not have a completely definite meaning. It may still express a range of possible meanings, differently interpretable by different participants or readers. This is very often the case at the point where it occurs. The context needed to specify its meaning very often at least partly follows its occurrence. So it may seem to have a more definite meaning retrospectively than it has instantaneously. In fact, depending on what follows, its meaning, as participants react to it, can be changed radically by what follows (retrospective recontextualisation). Analysing a text to see what is happening to meanings moment-to-moment yields a dynamical analysis; the overall net retrospective meaning when all is said and done, yields the synoptic analysis.

For a variety of good examples of rhetorical or speech act analysis, see Gee 1990 (esp. chaps 4 & 5); Green & Harker 1988; Grimshaw 1994; Lemke 1990a; Mann & Thompson 1988, 1992; Mehan 1979; Sinclair & Coulthard 1975. For discussions of evaluative and affective meaning, see Lemke 1988a, 1989b, 1990b; Martin 1992. For viewpoint analysis see discussions of heteroglossia in Bakhtin 1935; Lemke 1988a, 1995a, 1995b; and of social voices in Wertsch 1991. For dynamic and synoptic analysis see Lemke 1984, 1988b, 1991; Martin 1985, 1992; Ventola 1987.

Structural-Textural Analysis

Verbal data has social meaningfulness only as text, not as collections of isolated words or phrases (except statistically). How does a coherent, cohesive text differ from a random collection of grammatical sentences? How are texts and discourse events unified and subdivided into wholes and parts? How can we define the boundaries of a unit or episode of a text or verbal interaction? What binds the units of a text together?

Structural analysis of texts needs to be both "top-down" and "bottom-up", that is, it needs to consistently reconcile analyses that begin from the smallest units of meaning (normally phrases and clauses) and look for how these aggregate together into larger units, with analyses that begin from the largest units (normally activities and episodes or genres and their stages) and look for how these are composed of functional constituents. The largest unit of analysis for a spoken discourse text is the socially recognised activity-type in which the discourse is playing a functional part, or the smallest episode or subunit of that activity which contains the entire discourse event. A classroom Lesson is a typical activity-type of this kind. An episode of Going-Over-Homework or Working-in-Groups may form the more immediate context. The largest unit for a written text is normally the genre of which it is an instance, or the text itself (unless it is being analysed directly in its context of production or use, in which case the activity type applies, e.g. Lemke 1989c).

A genre is a text-type specified by identifying a common structure of functional units (obligatory and optional) that is repeated again and again from text to text. A speech genre is a generally a highly-specific activity-type accomplished mainly by verbal means. The term genre is more often used for types of written texts because they are more structurally standardised in our culture. A genre has a constituency structure in

which each constituent plays a functional role in the whole and has specific functional meaning relations to the other constituents on its own level. The largest units are often called stages, and they may be composed of smaller units, and these of still smaller ones, etc. Each constituent at each level of analysis should be defined in a way which is unique to the genre. A science lab report, as a written genre, might have major stages such as: Title, Author, Class, Statement of Problem, Description of Apparatus, Description of Procedures, Record of Observations, Analysis of Data, Conclusions, etc. The Description of Procedures might include a series of Procedure Statements, each saying what was done, when, and how. Each of these might not be composed of smaller genre-specific functional units, but only of grammatical units (i.e. the relationship changes from "composed of" to "realised by").

Some constituents of some genres have an intermediate level of organisation between genre-specific units and grammatical ones. These are often called rhetorical structures or formations (e.g. Lemke 1988b; Mann & Thompson 1988, 1992). They are found in essentially the same form in many different genres, but they have an internal functional or rhetorical structure in addition to the structure of their grammatical units. The most famous example in classroom discourse analysis is the IRF structure, typically realised as Teacher Question, Student Answer, Teacher Evaluation (see Lemke 1990a; Mehan 1979; Sinclair & Coulthard 1975). This could be considered genre-specific in classroom activity, but since it occurs in many different kinds of episodes (Lemke 1990a), it is more nearly a rhetorical formation. Something very similar occurs in courtroom discourse as well. More common and widespread examples include the simple Question-Answer pattern, or Examples-Generalisation, Event-Consequences, syllogisms, etc. They are, in effect, portable mini-genres.

Conversation analysis techniques often refer to particular cases as "adjacency pairs" (e.g. Sacks et al. 1974) Below the level of smallest genre-specific units and the moves within a rhetorical formation, we find the level of grammatical structure. Analysts should be aware that there are multiple simultaneous grammatical units structuring the same set of words, and that some of these may depend on intonation as well as word sequence. In Halliday's analysis, for example, a clause is simultaneously structured in terms of

Processes-Participants-Circumstances, Subject-Finite-Predicator-Complements-Adjuncts, Theme-Rheme, and Given-New.

Other models also have adopted the multi-structure approach (e.g. Chomsky analyses in terms of thematic roles, X-bar structures, and Logical Form). The boundaries of these different units are not necessarily the same. This brings us to the classic problem of textual structure: segmentation. Can a text be definitively divided at word boundaries into its constituent units at any level of analysis? The answer is: only sometimes. The same word can function as an element in different units, for different functions, on different scales. The boundary, particularly of a large, high-ranking unit (e.g. genre stage, rhetorical move) may be indeterminate in terms of lower level grammatical or word units because it is defined by several simultaneous criteria, each of which results in drawing the boundary in a slightly different place in the text.

As a general rule, units of meaning can have fuzzy boundaries in terms of units of form (or even in terms of units of meaning at a different level of analysis). Some texts are more rigidly structured than others. Some maintain, repeat, and complete particular genre patterns or rhetorical formations more consistently than others. Many texts frequently shift genre pattern or rhetorical strategy, with or without completion of those already started. Conversational discourse is notorious in this respect, but so are written texts by young writers who have not yet learned the genre conventions and borrow from the norms of conversational organisation. How do such texts maintain their coherence? In large part by topic continuity. More generally, by maintaining cohesion chains, whose members have no consistent structural-functional relations. If a structure looks like A-B-C-D, a chain looks like A-A-A-A.

Chains may be of many kinds. Lexical chains consist of words each of which may be the same word, have the same meaning in context, refer to the same referent, belong to the same semantic domain, etc. A short lexical chain may be accidental; a long one rarely is. Larger units than words may form chains, or strands. A structural pattern may be repeated (cf. rhetorical parallelism): A-B-C-D, A-B-C-D, A-B-C-D, etc. More commonly, and very importantly, a thematic pattern may be repeated, and varied, at different levels of abstraction (see Lemke 1995b for an extended analysis); this is very common in classroom teaching, and indeed in any text that has a lot to say about a small topic. Chains also normally interact with one another; that is, in each instance from two different chains, there is the same structural relation each time between the member of one chain and the corresponding member of the other. If we had a A-B structure, we could have an A-chain through the text and a B-chain, united by the fact that each A and B (or even just some of them) were connected in the A-B relation where they occurred in the running text (rather like a ladder). Not just chains of individual lexical items, but chains of whole thematic formations, can interact. It may take only a clause or nominal group (noun phrase) structure to tie members of two lexical chains together, but it can take much larger and more complex grammatical or rhetorical structures to do this between large thematic formations (see Lemke 1995b).

For further discussions of genre analysis see Bazerman 1988, 1994; Hasan 1984b, 1989; Martin 1989, 1992; Lemke 1988b, 1991; Propp 1928; Swales 1990; for activity-types see Lemke 1990a; Phillips 1983 (on participant structures); for rhetorical formations Lemke 1988b; Mann & Thompson 1988; for conversation analysis Sacks et al. 1974; for cohesive organisation, Halliday & Hasan 1976; Hasan 1984; Lemke 1988b, 1995b.

Case Studies and the Problem of Generalisability

How can verbal data and discourse analysis be used in studies of individual episodes and lessons, classrooms, and small groups? What is the value of such studies and how can we determine the generalisability of their findings?

Discourse analysis studies are often best when they examine a particular community in depth. Discourse analysis produces its greatest insights when rich contextual information can be factored into the analysis of each text or episode. For this reason, longitudinal designs or case studies are well suited for discourse analysis methods. Here we learn a great deal about a particular class, seeing repeated patterns within the data and a variety of strategies which create variations on those patterns. It is even possible to proceed to the level of individual psychological analysis, but in a larger sense all case

studies are about "individuals", and to a much greater extent than is true for other natural systems, human communities' individuality matters for the kinds of behaviors of the system that interest us.

It is not true that science should be only about generalised properties of classes of phenomena and not about unique properties of individual instances. The balance between these two approaches must be struck differently depending on the nature of the phenomena. Electrons seem to have no individuality that matters; biological systems do, but a great deal of their structure and behaviour remains constant for a species or variety. Developmental phenomena show a wide range of individual pathways, many of them approximately "equifinal", leading to the same end result (for example, language acquisition). Human communities and cultures are often more interesting for what is unique to them than for what they all have in common. Moreover, one of the important properties of any class is precisely the specification of how the members of the class differ from one another. Many sentences have a lot in common; that is the foundation of grammar. Many texts have a little in common, hence the concept of genre. But while the resources and strategies by which texts and discourse are constructed may be common to many texts, and help to specify how they may differ from one another, what is ultimately of interest about any text is its meaning, and that is its most unique feature. Discourse analysis will not tell us a lot about how all classrooms or all science writing is alike (it will tell us a little), but it provides us with the tools to analyse and understand what exactly is going on in any discourse or text we wish to analyse. That is as much as any theory really does for us in practice.

Protocol Analysis and the Problem of Interpretation

When task activities differ significantly from normal cultural routines, how will cultural patterns of language use be distinguishable from idiosyncratic constructions? What is the object of study that we construct from such data?

One important form of verbal data is generated when researchers construct special task activities that differ significantly from normal cultural routines. This follows the traditions of the natural sciences in devising tasks meant to reveal particular aspects of phenomena, but it encounters the risk (minimal for electrons and molecules, but already significant for organisms) that behavior under task conditions differs in important, and unknown, ways from that in normal routines. The essential context-sensitivity of meaning-based phenomena (meaning is selective contextualisation) strongly suggests that if we are interested in, say, a classroom phenomenon, that we study it in situ. If we supplement this with artificial tasks, it is then necessary to establish empirically that the differences between the task context and the natural context do not alter the phenomena of interest, or to identify in exactly what ways they do alter them.

Current models of situated cognition call into question the assumption that meaning-making processes can be assumed independent of local contexts, or even that "cognition" is a process in a system limited to the organism itself (as opposed to one that includes the organism's tools and the elements of the environment with which it interacts; cf. Lave 1988; Lave & Wenger 1991; Kirshner in press; Lemke in press-b).

Discourse analysis assumes that the resources and strategies (lexis and grammar, rhetorical formations, typical cultural narratives, genres, the principles of constructing

thematic formations, cohesion chains, etc.) used in producing discourse events and texts are characteristics of a community, rather than unique to an event in that community. They are part of its general cultural resources (and so differ from culture to culture and one community or subcommunity to another), but what it means to have a culture is that we preferentially deploy some of these resources in some contexts rather than others: that how we use the resources is essentially context-dependent. The analysis of the covariation between situational features and which lexical and grammatical resources are typically deployed in them is the subject of register theory (e.g. Gregory 1967; Gregory & Carroll 1978; Halliday 1977, 1978, 1991), which can also be adapted to analyse the clause-to-clause shifts in meaning that take place through a text (phasal analysis, e.g. Malcolm 1985). A similar theory for the deployment of the other resources does not yet exist (but see Kress & Threadgold 1988; Hasan 1994; Lemke 1995a; Thibault 1991). We should also recognise that while artificial activities may not be natural for the subjects asked to perform in them, they are in a sense a natural part of the culture of the researchers, which is thus mixed with that of the subjects by these procedures. There is perhaps a great deal to be learned about ourselves by analysing the nature of these tasks and their role in our own professional practice.

Comparative Studies and Cultural Bias

When we use discourse analysis and verbal data to compare males and females, middle and lower class subjects, widely differing age-groups, different cultural and linguistic groups, school practices and home, community, or professional practices, we necessarily introduce our own viewpoint, which is invariably closer to that of one of the categories compared than to the other's. Discourse data is not just sensitive to the context of immediate task and situation; it is also sensitive to the wider context of cultural norms and assumptions, knowledge, beliefs and values. The analysis of discourse data, its interpretation, is itself just more discourse, and discourse now from the point-of-view of the researcher's community.

Our research communities and their historical traditions are emphatically not equally balanced by gender, age, social class, or ethnic culture. Even studies which strive mightily for even-handedness and neutrality of description (e.g. Bernstein 1971, 1975; Hasan 1986; Heath 1983) are necessarily read by other researchers who will project their own values regarding what is better and what worse onto their descriptions of difference. In many other studies, even the questions which are asked of the data are asked from a very narrow range of human viewpoints. Discourse analysis is interpretation and it is viewpoint-dependent every bit as much as any other instance of discourse. The canonical procedures of discourse analysis which I have briefly sketched here provide a means for different analysts to systematically compare the many interdependent grounds of their respective interpretations. Whether they reach consensus or not is probably less important than the fact that the procedures be clear enough that others can enter into the discussion on common ground. These procedures, of course, are themselves the product of a narrow range of human viewpoints. We can hope that this range will widen as the field of discourse analysis, and our own society, matures toward more inclusiveness and respect for the value of diversity of viewpoints.

Curriculum Analysis and Evaluative Assessments: Ethical Issues

The methods of discourse analysis of verbal data can be used to compare curriculum documents, textbooks, and tests with classroom dialogue, teacher discourse, student writing, etc. They make possible rich descriptions of the lived curriculum, its relation to official curriculum plans, and the web of intertextuality among all the spoken and written language in which education is framed. They also make it possible to analyse how individual students use scientific language and concepts in a variety of situations, and to make this a basis for evaluative assessments.

These facts raise serious ethical questions regarding the appropriate use of discourse analysis methods in education. Our educational system operates within a larger social hierarchy of power and control. Administrative authorities seek to impose specific curricula on students, using teachers, textbooks, and the rest of the educational apparatus as the means. Their control is only as good as their means of assessing whether or not teachers teach, textbooks print, and students master what the curriculum mandates. From curriculum documents to test questions and responses, nearly all of this is in the form of text and discourse. Discourse analysis methods in principle allow far more precise ways of checking the match or mismatch of these elements than any other form of assessment or accountability.

Hopefully new assessment schemes will place more weight on practical skills and graphical modes of representation, but discourse forms will almost certainly still dominate (see Lemke in press for initial work on the analysis of visual representations and the role of mathematics in scientific discourse). For now, while automation of discourse analysis procedures remains thoroughly primitive, students and teachers who believe they have a right to control the content and directions of their own learning and teaching have little to fear from discourse analysis methods.

Meanwhile, the automation of new information technologies begins to offer at least more privileged students the opportunity to explore the universe of knowledge guided by their own evolving interests (cf. Lemke 1994, 1995c) rather than as prescribed by someone else's curriculum. Discourse analysis methods are already important in computer-based natural language systems for generating, analysing, and sorting texts. They will be even more important as components of the intelligent tutoring systems of the future, which will enable students to explore new information worlds more successfully with their help. Researchers of the next generation will help determine whether discourse analysis methods will be used to empower students in the new century, or more strictly control them.

Bibliography of Sources for Discourse Analysis Methods

Austin, J.L.: 1962, How To Do Things With Words, Oxford University Press, Oxford, UK.

Bakhtin, M.M.: 1935, 'Discourse in the Novel', in M. Holquist (ed.), The Dialogic Imagination (translation 1981), University of Texas Press, Austin, TX.

Bakhtin, M.M.: 1953, Speech Genres and Other Late Essays, (1986 edition), University of Texas Press, Austin, TX.

Bazerman, C.: 1988, Shaping Written Knowledge, University of Wisconsin Press, Madison, WI.

Bazerman, C.: 1994, 'Systems of Genres and the Enactment of Social Intentions', in A. Freedman & P. Medway (eds.), Rethinking Genre, Falmer Press, London.

Benson, J. & Greaves, W.: 1992, 'The Notion of Technicality in Register', in M. Davies and L. Ravelli (eds.), Advances in Systemic Linguistics: Recent Theory and Practice. Pinter, London, 205-220.

Bernstein, B.: 1971, Class, Codes, and Control, Volume 1, Routledge, London.

Bernstein, B.: 1975. Class, Codes, and Control, Volume 2, Routledge, London.

Bertin, J.: 1983, Semiology of Graphics, University of Wisconsin Press, Madison, WI.

Brazil, D., Coulthard, M., & Johns, K.: 1980, Discourse Intonation and Language Teaching, Longman, London.

Brown, G., & Yule, G.: 1983, Discourse Analysis, Cambridge University Press, Cambridge.

Cazden, C.: 1986, 'Classroom Discourse' in M. Wittrock (ed.), Handbook of Research on Teaching, 3rd edition, Macmillan, New York, 432--463.

Cazden, C.: 1988, Classroom Discourse: The Language of Teaching and Learning, Heinemann, Portsmouth, NH.

Coulthard, M.: 1977, Introduction to Discourse Analysis, Longman, London.

Coulthard, M. & Montgomery, M.: 1981, Studies in Discourse Analysis, Routledge, London.

Edwards, A.D. & Westgate, D.P.G.: 1994, Investigating Classroom Talk, Falmer Press, London.

Erickson, F.: 1982, 'Audiovisual Records as a Primary Data Source', Sociological Methods and Research 11, 213--232.

Erickson, F. & Shultz, J.: 1981, 'When is a Context?', in J.L. Green & C. Wallat (eds.), Ethnography and Language in Educational Settings, Ablex Publishing, Norwood, NJ, 147--160.

Firth, J.R.: 1957, Papers in Linguistics 1934--1951, Oxford University Press, London.

Gee, J.P.: 1990, Social Linguistics and Literacies, Falmer Press, London.

Green, J. & Harker, J. (eds.): 1988, Multiple Perspective Analysis of Classroom Discourse, Ablex Publishing, Norwood, NJ.

Gregory, M.: 1967, 'Aspects of Varieties Differentiation', Journal of Linguistics 3, 177--198.

Gregory, M. and Carroll, S.: 1978, Language and Situation, Language Varieties and their Social Contexts. London, Routledge.

Grimshaw, A.D. (ed.): 1994, What's Going On Here? Complementary Studies of Professional Talk, Ablex Publishing, Norwood, NJ.

Halliday, M.A.K.: 1967, Intonation and Grammar in British English, Mouton, The Hague.

Halliday, M.A.K.: 1976, Halliday: System and Function in Language, (G. Kress, ed.), Oxford University Press, London.

Halliday, M.A.K.: 1977, 'Text as Semantic Choice in Social Context' in T.A. van Dyck & J. Petofi (eds.), Grammars and Descriptions, de Gruyter, Berlin.

Halliday, M.A.K.: 1978, Language as social semiotic, Edward Arnold, London.

Halliday, M.A.K.: 1985, An Introduction to Functional Grammar, Edward Arnold, London.

Halliday, M.A.K.: 1991, 'Towards Probabilistic Interpretations' in E. Ventola (ed.), Recent Systemic and Other Functional Views on Language, Mouton/deGruyter, Berlin, 39--62.

Halliday, M.A.K. & James, Z.L.: 1993, 'A Quantitative Study of Polarity and Primary Tense in the English Finite Clause', in J.M. Sinclair, M. Hoey, & G. Fox (eds.), Techniques of Description, Spoken and Written Discourse, Routledge, London.

Halliday, M.A.K. & Hasan, R.: 1976, Cohesion in English, Longman, London.

Halliday, M.A.K. & Hasan, R.: 1989, Language, context, and text. Oxford University Press, Oxford, UK.

Hasan, R.: 1984, 'Coherence and Cohesive Harmony'in J. Flood (ed.), Understanding Reading Comprehension, International Reading Association, Newark, DE.

Hasan, R.: 1984b, 'The Structure of the Nursery Tale' in L. Coveri (ed.), Linguistica Testuale, Bulzoni, Rome.

Hasan, R.: 1985, 'Meaning, Context, and Text: Fifty Years after Malinowski', in J.D. Benson and W.S. Greaves (eds.), Systemic Perspectives on Discourse, Ablex Publishing, Norwood NJ.

Hasan, R.: 1986, 'The Ontogenesis of Ideology' in T. Threadgold, E.A.Grosz, Gunther Kress, & M.A.K. Halliday (eds.), Semiotics, Ideology, Language, Sydney Association for Studies in Society and Culture, Sydney.

Hasan, R.: 1989, 'The Structure of a Text' in Halliday, M.A.K. & Hasan, R., Language, Context, and Text, Oxford University Press, London.

Hasan, R.: 1994, 'The Conception of Context in Text' in M. Gregory & P. Fries (eds.), Discourse in Society, Functional Perspectives, Ablex Publishing, Norwood, NJ.

Hasan, R. & Cloran, C.: 1990. 'A Sociolinguistic Interpretation of Everyday Talk between Mothers and Children' in M.A.K. Halliday, John Gibbons, and Howard Nicholas (eds.), Learning, Keeping, and Using Language, John Benjamins, Amsterdam.

Heath, S.B.: 1983, Ways with Words, Cambridge University Press, London.

Hymes, D.: 1972, 'Models of the Interaction of Language and Social Life', in J. Gumperz & D. Hymes (eds.), Directions in Sociolinguistics, Holt Rinehart, New York, 35--71.

Kendon, A.: 1990, Conducting Interaction, Cambridge University Press, London.

Kirshner, D. (ed.): in press, Situated Cognition Theory, Erlbaum, Hillsdale, NJ.

Kress, G. & Threadgold, T.: 1988. 'Towards a Social Theory of Genre', Southern Review, 21, 215--243.

Lave, J.: 1988, Cognition in Practice, Cambridge University Press, Cambridge, UK.

Lave, J. & Wenger, E.: 1991, Situated Learning, Legitimate Peripheral Participation, Cambridge University Press, New York.

Lemke, J.L.: 1983, 'Thematic Analysis, Systems, Structures, and Strategies', Semiotic Inquiry 3, 159--187.

Lemke, J.L.: 1984, Semiotics and Education, Toronto Semiotic Circle Monographs, Victoria College, Toronto.

Lemke, J.L.: 1985, 'Ideology, Intertextuality, and the Notion of Register' in J.D. Benson, J.D. & W.S. Greaves (eds.), Systemic Perspectives on Discourse, Ablex Publishing, Norwood, NJ.

Lemke, J.L.: 1987, 'Strategic Deployment of Speech and Action, A Sociosemiotic Analysis' in J.Evans and J.Deely (eds.), Semiotics: 1983, Proceedings of the Semiotic Society of America `Snowbird' Conference, University Press of America, New York, 67--79.

Lemke, J.L.: 1988a, 'Discourses in Conflict: Heteroglossia and Text Semantics' in J.D. Benson & W.S. Greaves (eds.), Functional Perspectives on Discourse, Ablex Publishing, Norwood, NJ.

Lemke, J.L.: 1988b, 'Text Structure and Text Semantics' in R. Veltman & E. Steiner (eds.), Pragmatics, Discourse, and Text: Systemic Approaches, Pinter, London.

Lemke, J.L.: 1989a, 'Social Semiotics: A New Model for Literacy Education' in D. Bloome (ed.), Classrooms and Literacy, Ablex Publishing, Norwood, NJ, 289--309.

Lemke, J.L.: 1989b, 'Semantics and Social Values', WORD 40, 37--50.

Lemke, J.L.: 1989c, 'Making Text Talk', Theory into Practice 28, 136--141.

Lemke, J.L.: 1990a, Talking Science, Language, Learning, and Values, Ablex Publishing, Norwood, NJ..

Lemke, J.L.: 1990b, 'Technical Discourse and Technocratic Ideology' in M.A.K. Halliday, J. Gibbons, & H. Nicholas (eds.), Learning, Keeping, and Using Language, Selected Papers from the 8th AILA World Congress of Applied Linguistics, Volume II, John Benjamins, Amsterdam, 435--460.

Lemke, J.L.: 1991, 'Text Production and Dynamic Text Semantics' in E. Ventola (ed.), Functional and Systemic Linguistics: Approaches and Uses, (Trends in Linguistics, Studies and Monographs 55), Mouton/deGruyter, Berlin, 23--38.

Lemke, J.L.: 1992, 'Interpersonal Meaning in Discourse: Value Orientations' in M. Davies and L. Ravelli (eds.), Advances in Systemic Linguistics, Recent Theory and Practice, Pinter, London, 82--104.

Lemke, J.L.: 1993, 'Intertextuality and Educational Research', Linguistics and Education 4, 257--268.

Lemke, J.L.: 1994, 'The Coming Paradigm Wars in Education: Curriculum vs. Information Access' in Cyberspace Superhighways: Access, Ethics, and Control, John Marshall Law School, Chicago, 76--85.

Lemke, J.L.: 1995a, Textual Politics, Discourse and Social Dynamics, Taylor & Francis, London.

Lemke, J.L.: 1995b, 'Intertextuality and Text Semantics' in M. Gregory & P. Fries (eds.), Discourse in Society: Functional Perspectives, Ablex Publishing, Norwood, NJ, 85--114.

Lemke, J.L.: 1995c, 'Hypermedia and Higher Education' in T.M. Harrison and T.D. Stephen (eds.), Computer Networking and Scholarship in the 21st Century University, SUNY Press, Albany.

Lemke, J.L.: in press--a, 'Multiplying Meaning, Visual and Verbal Semiotics in Scientific Text' in J.R. Martin (ed.), Scientific Discourse, Longman, London.

Lemke, J.L.: in press--b, 'Cognition, Context, and Learning: A Social Semiotic Perspective', in Kirshner, D. (ed.), Situated Cognition Theory, Erlbaum, Hillsdale, NJ.

Malcolm, K.: 1985, 'Communication Linguistics: A Sample Analysis' in J. Benson & W. Greaves (eds.), Systemic Perspectives on Discourse, Volume 2, Ablex Publishing, Norwood, NJ.

Malinowski, B.: 1923, 'The Problem of Meaning in Primitive Languages', Supplement I to C.K. Ogden & I.A. Richards (eds.), The Meaning of Meaning, Harcourt Brace, New York, 296--336.

Malinowski, B.: 1935, 'An Ethnographic Theory of Language', Part 4 of Coral Gardens and their Magic, Volume 2, Allen and Unwin, London, 4--78.

Mann, W. & Thompson, S.: 1988, 'Rhetorical Structure Theory', Text 8, 243--281.

Mann, W. & Thompson, S. (eds.): 1992, Diverse Analyses of a Fund--raising Text, John Benjamin, Amsterdam.

Martin, J.R.: 1985, 'Process and Text' in J.D. Benson and W.S. Greaves (eds.), Systemic Perspectives on Discourse, Volume 1, Ablex Publishing, Norwood, NJ.

Martin, J.R.: 1989, Factual writing, Exploring and challenging social reality, Oxford University Press, London.

Martin, J.R.: 1992, English Text, John Benjamins, Philadelphia, PA.

Matthiessen, C.: 1992, 'Interpreting the Textual Metafunction', in M. Davies and L. Ravelli (eds.), Advances in Systemic Linguistics: Recent Theory and Practice, Pinter, London, 37--81.

Mehan, H.: 1979, Learning lessons: Social organisation in the classroom. Harvard University Press, Cambridge, MA.

Ochs, E.: 1979, 'Transcription as Theory', in E. Ochs & B. Schiefflin, Developmental Pragmatics, Academic Press, New York.

Poynton, C.: 1989, Language and Gender, Oxford University Press, Oxford, UK.

Propp, V.: 1928, The Morphology of the Folktale, (1968 edition), University of Texas Press, Austin, TX.

Sacks, H., Schegloff, E., & Jefferson, G.: 1974, 'A Simplest Systematics for the Organisation of Turn--taking for Conversation', Language 50, 696--735.

Scheflen, A.: 1975, 'Models and Epistemologies in the Study of Interaction' in A. Kendon, R.M. Harris, & M.R. Key (eds.), Organisation of Behaviour in Face--to--face Interaction, Mouton, The Hague.

Silverstein, M.: 1976, 'Shifters, Linguistic Categories, and Cultural Description' in K. Basso & H Shelby (eds.), Meaning in Anthropology, University of New Mexico Press, Albuquerque, NM.

Sinclair, J.M.: 1991, Corpus, Concordance, Collocation, Oxford University Press, Oxford, UK.

Sinclair, J., & Coulthard, M.: 1975, Towards an analysis of discourse, Oxford University Press, London.

Stubbs, M.: 1983, Discourse analysis: The sociolinguistic analysis of natural language, University of Chicago Press, Chicago.

Swales, J.: 1990, Genre Analysis, Cambridge University Press, Cambridge, UK.

Thibault, P.: 1991, Social semiotics as praxis, University of Minnesota Press, Minneapolis.

Tufte, E.: 1983, The Visual Display of Quantitative Information, Graphics Press, Cheshire, Connecticut.

Ventola, E.: 1987, The Structure of Social Interaction: A Systemic Approach to the Semiotics of Service Encounters, Frances Pinter, London.

Wertsch, J.: 1991, Voices of the Mind, Harvard University Press, Cambridge, MA.

Widdowson, H.: 1979, 'Rules and Procedures in Discourse Analysis', in T. Myers (ed.), The Development of Conversation and Discourse, Edinburgh University Press, Edinburgh.

Wilkinson, L.C.: 1982, Communicating in the classroom, Academic Press, New York.

Wortham, S.: 1992, 'Participant Examples and Classroom Interaction', Linguistics and Education, 4: 195--218.

Wortham, S.: 1994, Acting Out Participant Examples in the Classroom, John Benjamins, Amsterdam.

MULTIPLYING MEANING: VISUAL AND VERBAL SEMIOTICS IN SCIENTIFIC TEXT

J.L. LEMKECity University of New York

 

Multimedia Semiotics

Scientific research articles and other genres of formal scientific communication in print rely heavily on the use of visual representations such as graphs, tables, diagrams, and drawings as well as mathematical expressions. How are these symbolic presentations integrated with those made through normally textualized verbal language? How do we make meaning with such multimedia texts? What specific kinds of meanings have these multimedia genres evolved to help us make?

In this report on my current research-in-progress (Lemke 1993a, 1994), I would like to sketch a theoretical framework for investigating these questions and communicate some very preliminary findings. I will argue that human communication normally deploys the resources of multiple semiotic systems and combines them according to essentially functional principles. Scientific communication in particular seeks to make meanings that overflow the preponderantly typological principles of linguistic semantics and require their integration with the more topological modalities of visual semiotics and their extension through the hybrid resources of quantitative mathematics. I will also report the results of two preliminary surveys of the types and frequencies of non-textual presentations in formal scientific print communication and offer some semiotic analyses of the functional (presentational, orientational, and organizational) integration of text, tables, graphs, diagrams and drawings in these multimedia genres.

Social semiotics seeks to explicate how we make meaning with all the resources at our disposal: linguistic, pictorial, gestural, musical, choreographic, and most generally actional (e.g. Halliday 1978; Hodge & Kress 1988; Kress & van Leeuwen 1990; Lemke 1987, 1990, 1995a; O’Toole 1990, 1994; Thibault 1991). The most fully developed analysis so far is that for the deployment of linguistic resources in the construction of meaning in verbal text and discourse (e.g. Halliday 1985; Halliday & Hasan 1989; Lemke 1995a, 1995b; Martin 1992). But we never make meaning with language alone.

Meaning-making is always a material process as well as a social semiotic practice (Lemke 1984, 1993b, 1995a); as such it cannot be adequately understood in terms of any one semiotic modality, such as language. The acoustic soundstream of speech can not only be analyzed into linguistically distinctive features and units, which enable different voices to be heard as saying the same word, but in it we can equally hear those linguistically non-distinctive features that enable us to distinguish one individual voice from another, or a light-hearted from a melancholy speaking of the same words.

Languages, or linguistic semiotic resource systems, are analytical abstractions from embodied social practices: from material speakings and writings and the activities that provide the contexts on which their cultural meanings depend. Speech always co-deploys at least paralinguistic, and normally visual-gestural semiotic resources as well. Writing always deploys a visual graphological-typographical semiotic as well as that of language. In countless texts and acts of meaning-making the members of a community simultaneously use language and the semiotic systems of depiction and graphics, movement and gesture, music, fashion, food and every other mode of socially meaningful action in their culture.

Language co-evolved with other modes of meaning-making; it evolved to function in co-ordination with all aspects of our construction of meaning through semiotically mediated action and active perception. Speech co-evolved as part of interactional synchrony: the bodily and material integration of individual organisms into their ecosocial environments. The intonational patterns of speech and the musical patterns of song descend from common ancestral modes of behavior. The synchrony, not just in individuals but across dyads and groups, of verbal action with other body movements and rhythms signals the participation of gesture and movement in the unitary communication system from which we abstract the semiotic patterns we call language or gesture (cf. Kendon 1990, Scheflen 1975). Our perception as well as our production

of semiotic interaction makes use of visual and kinesthetic information and responsiveness as much as it does of the auditory channel.

In writing, ontogeny shows what phylogeny would tell us if it could: that writing and drawing also share common ancestry; that each generation must be taught anew to make the separation between them. Writing is not merely the annotation of speech, as drawing is not simply the inking of images. Drawing begins as the extension to paper of gesture, as the product of lasting visual traces of our gestures in acts which are indiscriminately gesturings and drawings (cf. Arnheim 1956, cited in Kress & van Leeuwen 1990: 25). And these are accompanied you can be sure by vocalizations which are not as distinct from other gestures as the abstractions of linguistics and the ideology of intellectual verbalism dispose us to believe. It is not surprising that children do as one act what adults have been taught to separate into two: drawing and writing (cf. Dyson 1991; Hicks & Kanevsky 1992). Like our first drawing, our first writing is not a representation of speech, but an extension of it that produces a lasting visual trace.

In the progressive cultural differentiation of the action-stream of communication and its lasting traces into speech, gesture, writing, and drawing there is a loss of unity, but there is also a semiotic gain. By a general process, common in development and evolution, and sometimes called semogenesis (Halliday 1975, 1992), when an original unity becomes semiotically construable as being composed of separable units, those units become free to combine in new ways, making new meanings possible, and the culture is then free to evolve new restrictions on these combinations, endowing them with yet another dimension of meaning (as "indices" of some situation-type or subcommunity; Lemke 1993b). When writing is distinguished from drawing, it becomes possible to compose multimedia texts, incorporating both semiotic modalities, and allowing us to multiply the meanings made with one by those made with the other, provided our community has established conventions for how to read such multimedia genres.

 

Multimedia Genres in Science

Natural science is preeminently a discourse about the materiality of the world. Because it is a discourse and a system of interdependent social semiotic practices and activities of many kinds united by discourses, science works to bridge the gulf between the analytical meaning systems of our culture and the underlying material processes by which we as organisms interact with and are integrally part of a physical, biological, ecological world. The materiality of physical processes is what overflows analytical categorization, for unlike semiotic reality, physical reality is not defined by difference but by interaction. Semiotic systems like language make meaning by combining differences, by making analytical-categorial distinctions and then distributing distinctive sign-bearers in non-random ways. But there is another kind of meaningfulness in which we participate through our material being: the interactions with environments that make us what we are, that enable us to live, grow, develop, and surprise ourselves. (This model of material semiotics and ecosocial systems is developed further in Lemke 1995a.)

In its efforts to describe the material interactions of people and things, natural science has been led away from an exclusive reliance on verbal language. It has tried to find

ways to describe continuous change and co-variation, in addition to categorial difference and co-distribution. It has tried to describe what we know through our perceptual Gestalts and motor activities, to construct representations of the topological as well as the typological aspects of our being-in-the-world. Language and other typologically oriented semiotic systems have not evolved in this way. Language is unsurpassed as a tool for the formulation of difference and relationship, for the making of categorical distinctions. But it is much poorer (though hardly bankrupt) in resources for formulating degree, quantity, gradation, continuous change, continuous co-variation, non-integer ratio, varying proportionality, complex topological relations of relative nearness or connectedness, the interpenetration of different dimensionalities, or nonlinear relationships and dynamical emergence.

But all these things can be enacted operationally through other forms of human motor activity, in the laboratory, or just in our normal biological ways of interacting with our environments (cf. Lemke 1996). We can indicate modulation of speed or size, or complex relations of shape or relative position, far better with a gesture than we can with words, and we can let that gesture leave a trace and become a visual-graphical representation that will sit still and let us re-examine it at our leisure. The dynamical aspect lost in the static graph is recovered in the visual simulation or animation, which we can repeat at will, speed up or slow down and freeze in time. Visual representations present events, happenings, doings in ways which retain their topological and dynamical character. Verbal representations do not do this. Vocal representations, as in speech performances, do retain aspects of the flow, the modulation of pacing and intensity of events (cf. Hymes 1981; Bourdieu 1972, 1990), but they do so through the motor dimensions of speech, not its specifically linguistic ones. (For related arguments see Bateson 1972, or in Wilden 1980, on the relations of "analogue" and "digital" communication, Thom’s 1975 discussion of "structural stability" and "morphogenesis," or Peirce’s 1955 discussion of iconically and indexically based semiosis vs. symbolic meaning.)

When we investigate the ebb-and-flow of material phenomena, we not only need to represent them but also to analyze our representations. Our tools of analysis are the predominantly typological ones of language and the semiotics of categorical distinction in other modalities. The system of discourses and allied practices that constitutes science as a social network of activities and actors (Latour 1987) has historically evolved a new semiotic modality to bridge betweent the typological and the topological: mathematics. Mathematics has its origins in natural language, but everywhere that natural language was semantically weak for purposes of representing the features of material processes of practical interest to our ancestors, mathematics extended its categories (non-integer numbers and ratios, quantitative relationships, variables, functions, infinitesimals, topologies, etc.) and made possible the kinds of meanings that could have one foot in each camp. They could vary continuously like visual and motor processes and graphical representations, but they could also be integrated into verbal text and discourse in ways these could not.

Science is not done, is not communicated, through verbal language alone. It cannot be. The "concepts" of science are not verbal concepts, though they have verbal components. They are semiotic hybrids, simultaneously and essentially verbal-typological and mathematical-graphical-operational-topological. The actional, conversational, and written textual genres of science are historically and presently, fundamentally and

irreducibly multimedia genres. To do science, to talk science, to read and write science it is necessary to juggle and combine in canonical ways verbal discourse, mathematical expression, graphical-visual representation, and motor operations in the "natural" (including human-as-natural) world.

When scientists think, talk, work, and teach (cf. Lemke 1987, 1990; Ochs et al. in press -a, -b) they do not just use words; they gesture and move in imaginary visual spaces defined by graphical representations and simulations, which in turn have mathematical expressions that can also be integrated into speech. When scientists communicate in print they do not produce linear verbal text; they do not even limit their visual forms to the typographical. They do not present and organize information only verbally; they do not construct logical arguments in purely verbal form. They combine, interconnect, and integrate verbal text with mathematical expressions, quantitative graphs, information tables, abstract diagrams, maps, drawings, photographs, and a host of unique specialized visual genres seen nowhere else.

I have recently made some preliminary surveys of professional scientific print publications to sample the relative frequency and common types of non-verbal-textual semiotic expression. Here are some of the results:

Survey I

23 articles and chapters (341 pages in total):

All feature articles from Bull NY Acad Med 68(3) 1992(4 articles on biomedicine, 1 on urban language diversity)

1 research off-print on cellular development All chapters from a conference proceedings on applications of

chaos theory in ecology, medicine, astronomy, physics 1 book chapter on ecological succession (advanced treatise) 1 lengthy journal review article on thermodynamics 1 advanced textbook chapter on nonlinear dynamics

Average length 14.8 pp.

Average graphics 16.2 per articleAverage per page: 1.1Maximum 4.3 per page.

Articles with equations, 17.Average length 14.9 pp. Average equations 20.8 per articleAverage per page: 1.4Maximum, math intensive, 3-4 per page.

Counted as graphics were all figures, tables, charts, graphs, photographs, drawings, maps, and specialized visual presentations. Mathematical equations were not counted in this total and are reported separately in these statistics. Only mathematical expressions set off typographically from running text were counted. Mathematical expressions are often directly integrated into a line of running verbal text, and into its syntax, which may be extended in register-specific ways beyond the syntactic possibilities of other registers.

The clear finding here is that there is typically at least one and often more than one graphical display and one mathematical expression per page of running text in typical scientific print genres. There can easily be 3-4 each of graphics displays and mathematical expressions separate from verbal text per page. (It is actually unusual to find high concentrations of both equations and graphics in the same article or on the same page. Experimental-empirical reports tend to have more graphics, theoretical analyses more equations.)

Survey II

Part A. All 20 articles from 2 issues of Physical Review Letters (62 pages total)

Average length: 3.1 pp

Total graphics, 76. Average per article, 3.8; Average per page, 1.2Maximum 2-3 per page.

Total equations, 169. Average per article, 8.5Average per page, 2.7Maximum, in theoretical articles, 6-7 per page.

Part B. All 31 Technical Reports from 2 issues of Science (74 pages total)

Average length: 2.4 pp.

Total graphics, 187Average per article, 6.0Average per page, 2.5

Total equations, 13[only 3 articles used equations, 4.3 per article, 1.9 per page]

Total tables, 39Average per article, 1.2

Total graphs, 30Average per article, 1.0

Visuals (excludes math and tables):

Average per article, 4.7 Average per page, 2.0

Physical Review Letters is probably the single most prestigious journal in the physical sciences. It accepts only brief reports of major new findings (experimental or theoretical) for rapid publication. Each typical 3-page article integrated four graphical displays and eight set-off mathematical expressions. Some had as many as three graphical displays per page of double-column text, or as many as seven equations per page.

Science is, alongside its British counterpart Nature, the corresponding journal for the biological and earth and space sciences. Each of these publishes longer feature articles as well as news items, all of which were excluded to achieve comparability to PRL. Mathematical expressions were less frequent in Science (only 10% of articles, all of them in non-biological subjects, included any equations), but visual presentations other than Tables, such as photographs, molecular diagrams, etc. were more common. A typical page has two non-tabular visual-graphical representations integrated with the verbal text. Each short (2.4 page) article typically has six graphics, including at least one table and one quantitative graph.

These were very preliminary surveys. There are considerable difficulties in deciding in practice how to count visual and mathematical presentations (count all lines of mathematics, or only separately numbered equations? count all graph lines in the same figure or only the figure as a whole? count combined tables separately or as a unit?). In all cases the statistical estimates given are conservative in the sense that almost any other counting choices would increase the frequency of mathematical and visual-graphical elements per article and per page.

To appreciate the absolutely central role of these non-verbal textual elements in the genres being characterized, it may help to ponder a few special (but hardly unique) cases:

In one advanced textbook chapter, a diagram was included in a footnote printed at the bottom of the page (Berge et al. 1984: 84). It was necessary for even the minor point being made there. (Figure 4)

In one 7-page research report in Nature, 90% of a page (all but 5 lines of main text at the top) was taken up by a complex diagram (see Figure 5, from Svoboda et al. 1993) and its extensive figure caption.

The main experimental results of a 2.5-page report in Nature were presented in a set of graphs occupying one-half page and a table occupying three-fourths of another (see Figure 2 , from Martikainen et al. 1993). The main verbal text did not repeat this information but only referred to it and commented on it.

In most of the theoretical physics articles, the running verbal text would make no sense without the integrated mathematical equations, which could not in most cases be effectively paraphrased in natural language, even though they can be,

and are normally meant to be read as if part of the verbal text (in terms of semantics, cohesion, and frequently grammar).

Presentational, Orientational, and Organizational Meaning

How are we to analyze the ways in which verbal text and mathematical and visual-graphical media combine in scientific genres? Within what theoretical framework can we examine how people make meaning by linking text and figure, sentence and equation, caption and photo? I have already suggested that the possibility of functionally integrating language, drawing, diagrams, pictures, and mathematics arises from their common origins in material human communication processes and their long co-evolution with one another in the uses we have historically made of them in our culture.

More specifically, I believe that all meaning-making, whatever semiotic resource systems are deployed, singly or jointly, has become organized around three generalized semiotic functions (Lemke 1989, 1990, 1992, 1995). When we make meaning we always simultaneously construct a "presentation" of some state-of-affairs, orient to this presentation and orient it to others, and in doing so create an organized structure of related elements. (For more detailed discussion of these presentational, orientational, and organizational functions, see below.) When we do this, as nearly as is possible, entirely with the resources of language, then these functions correspond to the ideational, interpersonal, and textual linguistic "metafunctions" of Halliday (1978). When we do this in "talking science" (Lemke 1990) or in any other specialized linguistic register (Gregory 1967, Halliday 1978), we deploy some of the resources for each function with greater and others with lesser frequency (Halliday 1991). Moreover, as Martin (1991, 1992) has emphasized, we form particular, culturally typical combinations of meanings about things, attitudes toward them and our addressees, and ways of organizing our language, which in effect constitute the genres of a discourse community (see also Hasan 1995, Lemke in press -a,-b; Bazerman 1988 for additional views of the relations of genres and social meaning).

Let’s consider just a bit more of what we know about functional integration in linguistic meaning as a background for the more general case of multimedia semiotics. Linguistically we have evolved more than just functional combinations, we have evolved typical discourse formations and stylistic strategies, which give these combinations a sense of familiarity even when their content is novel (cf. Kress & Threadgold 1988; Lemke 1988, 1991, 1995a, in press -a, -b). We use texts already in circulation, and discourse resources of the order of text (typical stories, typical patterns of scientific explanation, typical progressions of metaphors) as resources for making meaning, every bit as much as we use more formal, paradigmatic, system-like resources such as lexicogrammar or the semantic options for speech acts.

A clause or sentence means not just through the selections it makes in these systems, but through its place in a larger text organization and its intertextual relations to other texts (Lemke 1985, 1988a, 1988b, 1995a, 1995b). It means something not just through its own selections, but through how those selections continue or develop the selections of prior clauses (and may be reread in light of subsequent ones; cf. Halliday 1982, Lemke 1991). Moreover, ideational-thematic (presentational) meanings contribute to organizational-textual ones (e.g. through cohesion and cohesive harmony, cf. Halliday

& Hasan 1976, 1989; Hasan 1984; Lemke 1988b, 1995b) and so do intepersonal-attitudinal ones (cf. Lemke 1992; Thibault, in press, on global modal programs). Likewise, organizational meanings contribute to ideational ones by defining clauses, clause-complexes, rhetorical structure units (e.g. Lemke 1988b, 1990, in press -a, -b; Mann & Thompson 1986), and larger units of intratextual relationship, and in similar ways to the interpersonal and attitudinal texture of a text. Ideational choices of lexis, in turn, inevitably contribute to the attitudinal stance of a text to its audience, to its content, and to other text-embodied viewpoints. Conversely, we read the ideational meaning of a text differently depending on our interpretation of its ironies and other attitudinal features.

In these ways and many others, the meaning of the linguistic text is the product of the subtle, conventional and creative interplay of presentational, orientational, and organization aspects of meaning-making. How much more so is this true when we take into account the co-deployment of visual, gestural, graphical, mathematical, and other resource systems! Each of these semiotic systems provides resources specialized for each of the three generalized semiotic functions. When the resources of multiple semiotic systems are codeployed:

1. Each semiotic can contribute componentially to each each functional aspect of meaning (e.g. lexical name and figural image to a presentational construct),

2. Each can internally cross-modulate meanings across functional aspects (e.g. alternating point-of-view shots helping to construct a visual organizational sequence), and

3. Functionally specialized meaning resources in one semiotic combine with those for a different function in another semiotic to modulate any aspect of the meaning of the joint construct (e.g. the visual juxtaposition of verbal captions can allow their thematic meanings to interact with the different sizes of two graphs to determine their relative importance).

In multimedia genres, meanings made with each functional resource in each semiotic modality can modulate meanings of each kind in each other semiotic modality, thus multiplying the set of possible meanings that can be made (and so also the specificity of any particular meaning made against the background of this larger set of possibilities).

This combinatorial semiotic principle provides not just a theoretical framework, but an analytical engine for investigating multimedia semiotics. It poses for us the possibilities and the questions, what to look for and where to look. We will find some of the possibilities realized in some genres and not in others, in some texts perhaps and not in others. Just as the notion of grammatical metaphor (Halliday, 1985, this volume) shows us how our meaning-making potential is enlarged when verbal resources normally used for one function (e.g. nouns for Things) are deployed also for another (nouns for Processes, which thereby become semiotically both in some ways Thing-like while in others still Process-like, a new semiotic hybrid reality), so the principle of functional cross-multiplication in multimedia genres shows us how we can mean more, mean new kinds of meanings never before meant and not otherwise mean-able, when this process occurs both within and across different semiotic modalities (i.e. language, visual representation, mathematics, etc.).

Because of their importance to the analysis, let me specify the three generalized semiotic functions a little further:

Every meaning-making act constructs a presentational "state-of-affairs" that construes relations among semiotic participants and processes as if they were being observed, objectively and synoptically, from some outside vantage point. In language, this is the so-called representational or propositional function (Halliday’s "ideational" or "experiential" metafunction, for which I also use the term thematic). It defines the sense in which we speak "about" something, construct a theme or topic, make predications and arguments. In less philosophical and more linguistic terms, this means that we deploy the resources of grammar and lexis to specify some process or relationship and its semantic participants (agents, patients, instruments, etc.) and circumstances (when, where, why, how, under what conditions, etc.) In visual depiction, this is the figural or representational function that presents to us a scene whose elements we can recognize and which have comprehensible relations to one another in terms of the typical scripts of that scene. This is what tells us what we are being shown, what is supposed to be "there", to be happening, or what relations are being constructed among the elements presented.

At the same time every meaning-making act constructs an orientational "stance" toward that state-of-affairs, often to indicate how true or certain the producer wishes the interpreter to take it as being, or to indicate an evaluation of it as good or bad, ordinary or suprising, necessary or obligatory, in the perspective the producer is creating for the interpreter. But orientational meaning goes naturally beyond this to also include the construction of a social relationship between producer and interpreters (present or imagined), and more generally a relative positioning of the producer and "text" (i.e. semiotic production or multimedia object) in the whole social space of possible discourses and viewpoints on the state-of-affairs. In language, this is the pragmatic or interactional function (Halliday’s "interpersonal" metafunction, including "attitudinal" meanings), through which we take a role in the communication event and construct the nature of the "speech act" we are performing (informing, querying, commanding) and the social relationship to the addressee we are enacting (bullying, beseeching, promising, threatening). This is always done in the context of larger social relations and groupings that transcend any particular communicative event or text, and against the background of the various other texts in the community constructed from other possible points of view (cf. Lemke 1985, 1993c; Bakhtin 1935/1981 on heteroglossia, and commentary in Lemke 1988a, 1995a).

In depiction, every image takes an orientational stance (O’Toole’s 1994 term is "modal") which positions the viewer in relation to the scene (e.g. intimate, distant; superior, subordinate), establishes some sort of evaluative orientations of the producer/interpreter toward the scene itself (tragic, comic; normal, surprising), and does so against the background of other possible viewpoints and depictions of similar scenes.

Finally, every meaning-making act constructs a system of organizational relations defining wholes and parts of those wholes, both in the semiotic space of the text and in the (ecosocial) interactional space of the meaning-making act itself. Language creates words-in-phrases, phrases-in-clauses, chains of reference and cohesion, and larger and subtler structures and textures of the verbal text (Halliday’s "textual" metafunction, cf. Halliday & Hasan 1976, 1989; Lemke 1988b, 1995a; Matthiessen 1992). Conversation

as an activity creates in-groups and out-groups, and shifting dyads within larger groups. Depiction deploys compositional (O’Toole 1994) resources to organize the visual text into elements and regions, and to link disjoint regions by such features as color and texture. As material objects, depictions participate in interactions that define parts and unite them into wholes in the ecosystem networks where objects are viewed and used.

In recent work, Michael O’Toole (1994) and, independently, Gunther Kress and Theo van Leeuwen (1990) have demonstrated these homologous dimensions of meaning in visual semiotic productions of many kinds: paintings, sculpture and architecture; photographs, advertisements, and cinema. Kress and van Leeuwen (1990) have also considered some of the parallels between text and drawings in children’s writing. Their work clearly shows the usefulness of taking a "trifunctional" perspective such as that sketched here in visual semiotics, and the interesting similarities and enlightening differences between the specific meaning systems of these analytically distinct, and functionally and culturally specialized, semiotic resource systems and the genres we produce by deploying them.

In an early effort (Lemke 1987) I analyzed how science teachers and students made sense with each other by codeploying verbal, gestural, and pictorial resources. I found that if we regard each of these as constituting a separate "channel" of communication, then sometimes the same or equivalent information passes nearly simultaneously in more than one channel, sometimes the information in the two channels is complementary, and sometimes information comes first in one channel, and later in another. It became very clear to me that the meanings that were being constructed were JOINT meanings produced in the intersection of different semiotic systems. While it was useful to analytically separate these into different "channels," there was also an underlying unity to the meanings produced. Their separation neglects this fundamental unity of communicative meaning-making which makes the co-ordination among channels not only possible, but normal.

Genres within Scientific Text

How can we read scientific media that combine verbal text with mathematical expressions and various visual-graphical presentations? What are the component genres within such scientific texts? How are they typically integrated with one another? How, across semiotic modalities, are resources for presentational, orientational, and organizational meaning combined? Here I can only begin to point at some interesting phenomena that warrant further investigation and study.

The medium of printed scientific texts is first of all a visual one. Even the linguistic meanings are presented through the visual semiotics of orthography and typography, including all matters of page-layout as well as choices of font style and typeface sizes, the use of headings and footers, etc. It is precisely because language here is present through a visual semiotic that it is so readily integrated with other systems of visual meaning. (In face-to-face communication, where language is present in a motor-acoustic medium, it is our "vocal gestures" that are integrated with other motor gestures, some being more perceptible auditorily, others more so visually. Cf. Scheflen 1975).

Typography is quite conventionally used as an orientational as well as an organizational resource in printed text. Orientationally, the use of italic and boldface types signals

emphasis or importance, as does the relative point-size of type in titles, headings, abstracts, footnotes, captions, labels, etc. Organizationally, paragraphing and sectioning of text, and geometric relations of figure space to caption space indicate to us which elements are to be preferentially read in relation to which other elements; what goes with what.

Scientific text is not primarily linear, it is not meant to be read according to a unique implied sequence and represents a primitive form of hypertext (cf. Lemke 1993d). Footnotes represent an optional branch for readers, so do figures and their captions, and the parenthetic or main-text expressions such as "(Table 3)" or "as seen in the first table" which point to them. Spoken language is linear in this respect, but no visual semiotic can be, for all are at least two-dimensional and any one-dimensional sequence represented in two-dimensions can be accessed at any point at any time. Many scientists, after reading the title and abstract, may skip to the end to see an endnote telling who supported the research or to skim the references for familiar or unfamiliar citations. They may well read the tables or graphs first, and then their captions, and only then the main text. Some may pay more attention to the equations than to the words in which they are embedded. These are the habits of expert readers, those who could themselves have written this text or one very like it.

The principal mini-genres of the scientific research article, in addition to its Main Text, include: Title, Author List, Author affiliations or addresses, Abstract, Section Headings (which may be considered part of the Main Text), Figures, Figure Captions, Tables, Table Captions, Priority dates, Reference List, Footnotes/Endnotes, and Acknowledgments. Figures, as a typographical component, are regions of the page which are not set in type but produced by some more photographic or direct image-printing process. They include: Photographs, Drawings, Diagrams, Graphs, and Maps. Tables, by contrast, are set in type, and are the most text-like of the non-textual visual presentations. Almost all Figures have labels as well as captions, and often also brief textual notes printed as part of the Figure proper (legends).

Tables as Textualizable Visual Displays

Tables have their most direct origin in normal written text. They carry textual ellipsis to its greatest extreme, using visual organizational resources to enable meaning relations to be recovered from bare thematic items in the absence of grammatical constructions. There is always, however, an implied grammar, and a recoverable textual sentence or paragraph for every table. In my work on intertextual thematic formations (Lemke 1983, 1985, 1988, 1990, 1995b) originally for scientific discourse, I noted that listeners and readers are expected to be able to supply the canonical semantic relations of thematic terms, which are often underspecified or omitted. This is done by familiarity with a canonical pattern of semantic relations (the thematic formation), either from another text or discourse, or from an earlier section of the present one.

Scientific tables most often have numerical entries, though not always, and the grammar of the numerical entries is generally that of the normal nominal group, with the Head constant through a row or column of the table and written at its end, and the table entry functioning as numerative of that Head. Table I, however, shows a purely verbal table, with a brief introductory text, serving as a summary conclusion of a conference paper (Percival 1987). In the conference presentation it would almost certainly have been

projected as a slide. Having read the text of the paper (excepts below), the reader would be in possession of a thematic formation whose slots could be readily filled by the pairs of entries in this two-column table. If it were glossed as full text, it might read something like this:

"In the old view of Hamiltonian dynamics [lefthand entry] was a key concept, but in the new view it has been replaced by the idea of [righthand entry]."

CONCLUSIONS

The complete change of viewpoint described earlier can now be summarized, as in table 1, showing the old ideas, and how they have been modified in the last few decades.

TABLE 1. CHANGING VIEWPOINTS IN HAMILTONIAN DYNAMICS

Old view New viewintegrability invariant setsstability and instability of orbits

regular and chaotic motion

variational principles for orbits

variational principles for invariant sets

convergence of perturbation series

K.A.M. theorem and analyticity in domains

continuity fractals and renomalization

analysis geometry, computers, and number theory

determinism probability

[From Percival 1987: 142; reproduced by permission]

As is typical with thematic formations, there are many possible textualizations, but the underlying semantic relations among thematic items remain the same. Here is one instance of the formation from the text of the article which is being summarized in the Table:

According to one point of view, expressed by Laplace, dynamical systems are completely deterministic, so probability theory can have no relevance. But this point of view requires [an impossibility]. ... For ... [certain] systems [modern] theorists have

introduced symbolic dynamics, which shows precisely how probabilities enter into deterministic dynamics. (Percival 1987, p.139-140)

It corresponds to the last pair of entries in the Table.

When numerical values are placed in a Table, it is still meant to be read, at least in principle, as full verbal text. While this convention today is so well-established that it is not in evidence in the surface forms of most Tables, a historical example shows the textual nature of Tables more clearly.

In a 1734 edition of an early treatise by Johan Bernoulli (De motu musculorum, p.464), a Table is set out giving the values of various lengths and angles resulting from a calculation of the movement of an animal’s jointed appendages subject to a certain weight, set at a certain angle (see Figure 1). There is an accompanying Figure of course to label these parameters. In the Table, we find (left-to-right) first a thin

vertical column with the word "Positis" written vertically, then two columns of numbers headed "Arcu" and "Longitudine" and written in the usual horizontal way, followed by another very thin vertically ruled column containing the single word "erit" (lower case) written vertically again, and then a series of further columns of numbers headed "Radius" "Sinus" etc.

 

 

 

Figure 1. Table from Bernoulli (1734: 464)

[Reproduced courtesy of the New York Academy of Medicine Library]

 

Here we have the grammatical elements of the complete Latin sentence which can be constructed for any row of the Table: Given [Positis] arc ..., longitude ..., [then] radius,

sine, etc. will be [erit] ..., ..., etc. Latin is already more sparing lexically than modern English, and there may certainly still be some ellipsis (of aequale and perhaps of et for "equal to" and "and"), but the Table has been set with the most necessary words, especially the future tense of the copula erit, to make the Table a representation of verbal text. (The inflections on Positis, arcu, longitudine are all ablative case, appropriate to textual grammar, as opposed to the nominatives usual for isolated headings.)

In a modern Table (Figure 2, from Martikainen 1993), we find a mixture of textual entries and numerical ones, including the "Tree Stand" entries which preserve the nominal group form. Nowhere in the Main Text of this article do we find a sentence that textualizes any row of this Table, but no competent reader of the text would be unable to do so. What is important to notice here is that the Table makes sense precisely because it is textualizable, i.e. because there is a thematic formation in this community, intertextually available, which enables readers to make specific semantic relations between "4.3" and "pH", "Drained" and "Virgin", "Tall Sedge Fen" and "Total N in peat".

Figure 2. Table 1 from Martikainen et al. (1993: 52)[Reproduced from Nature, by permission]

The important visual resources here, apart from the orthography of words, numbers, special symbols and abbreviations, are organizational ones, arranging in vertical columns and horizontal rows the semantically homogeneous (and therefore contrastable,

comparable) items, and the semantically heterogeneous (and therefore combinable, presentationally interdependent) items, respectively. We can read the Table either for information about a site, for comparison of sites, to see the range of variation in a site-feature across sites, or to look for co-patternings. The topological meanings represented by decimal fractions and percentages as measurements of continuously varying physical quantities have been integrated visually as well as implicitly linguistically with the typological categories of this register. (Note, by the way, how the semantics of standard grammar is extended in delicacy by such quantitative expressions as "a tree stand of 50 cubic meters of wood per hectare" or "a rate of emission of nitrogen oxide of 0.143 grams of nitrogen per square meter per year.")

A relevant bit of the Main Text reads: "... the higher tree growth in the minerotrophic sites [fn8] (Table 1) had increased evapotranspiration and contributed to the lower water table (Fig. 1a-c)." In this one compound clause we are referred to a citation for the published details on the sites, to the Table for the "higher tree growth" at the "minerotrophic sites", and to the Figure for graphs of water table levels over time. Even when we refer to Table 1, we need to look to a footnote within the Table caption, hung on the Header text "Site number and type [*]", to identify which sites are "minerotrophic," and recognize this as a classifier in contrast with "ombrotrophic." The visual organizational conventions here are critical to following the verbal presentational argument. We could even pick up the -trophic contrast by noting the positions of these words as Headers for two columns of graphs in Figure 2. The graphs also show "error bars" indicating the reliability or warrantability, and so probability of error, in the data: a visual presentation of a mathematical formulation of an orientational meaning.

Abstract Graphs and Diagrams

It is not very far historically from a Table with numerical, or repeated symbol, entries, to a primitive bar-graph, and from there to a line-graph. Even Cartesian line-graphs and their corresponding mathematical functions can to a certain extent be "read" as text insofar as the axes have numerical scales, allowing the graph to be read as a geometrical condensation of a Table like that in Bernoulli, "when x has the value ..., y will have the value ...." This is certainly true for data-graphs, but even here, visual semiotics begin to allow us to see "trends" and "functional dependencies" in the data, which no textualization of it as information would lend itself to so directly. We apprehend the "patterns" in the data when displayed as a graph differently than we do when it is displayed as text, or even as a Table.

Such patterns are important in the value-scheme of natural science. Mathematical functions abstract these patterns from the data and even the variables, the thematic items, the measured phenomena among which they are the patterns. In scientific theory, patterns are generated from assumptions. Initially these are verbal-semantic statements, then they are mathematical expressions and/or graphical images or imagined or displayed dynamical simulations, but always primarily of quantitative patterns and covariations, the very things language has not evolved to present very well. So, in addition to mathematical expressions, scientific practice makes use of abstract graphs which only show conceptual relations, and not actual data. Here we leave behind any possibility of textualization as for Tables, but now there is a new sort of textualization possible, by way of mathematics. The abstract graph has a mathematical formulation,

and that in turn is rooted in an extension of the semantics of natural language and its grammar.

Abstract graphs are more nearly pictures than text. Or they sit about mid-way. If we begin from the other end of the typology (text and verbal semantics) to topology (continuous variation and interaction) line, we must start from our visual perception of real objects, and then their holographic or photographic representation. A photograph can be abstracted into a drawing (with third-dimensional perspective or without), and from the drawing it is only a few more steps to the abstract diagram, which is still interpreted as if it were a drawing or a photograph or a visual percept. For each step we foreground some features as more important, as more significantly in contrast with potential alternatives and so more in need of specification, or else by mere convention we let the part stand for the whole, we suggest a whole, which the reader must supply, fitting intertextual (inter-pictorial?) information to the presumptive pattern. This is not at all unlike the corresponding process with thematic formations for verbal meaning.

Abstract diagrams and abstract graphs are sufficiently close in their visual semiotics that they can be directly combined visually and read as two parts of the same visual presentation. As an example, consider Figure 3 (from Berge et al. 1984: 84). The fine horizontal dotted lines visually integrate (organizational function) the abstract graph at right with the abstract diagram at left. This is possible here because one of the variables, z, in the graph is a measure of spatial position, and the scale of the graph has been set to correspond exactly to the visual representation of the same spatial distance in the diagram. But otherwise we are still in two visual worlds on the two sides of this Figure. At the left the bold arrow of "g" points down to indicate the pull of gravity and the horizontal dimension is still spatial. At the right, "z" on its vertical axis is only an arbitrary, scaleable measure of spatial position, oriented to "real space" only by the visual connection here (and by the conventional use of the co-ordinate name "z" for vertical positioning); the horizontal direction now represents the value of the temperature of places with vertical position "z", it is not spatial at all, except metaphorically.

Figure 3. Figure V.3 from Berge et al. (1984: 84)[Reproduced by permission]

On the same page of this text appears footnote 58 (Figure 4), which is both textually linked to, and visually echoes, the diagram portion of Figure 3. The diagram is rotated by 90 degrees, exchanging vertical for horizontal and vice versa. This change would not matter at all to the meaning of an abstract graph, but in this case the diagram is once again oriented on the page to "real space" with the force of gravity still acting down, even though the meaning of the downwards arrow is now ambivalent (the footnote text refers to both the fall of the liquid and to the force, gravity, which produces that fall). The full meaning of footnote 58 requires an integration of presentational, orientational, and organizational contributions from its diagram, the footnote text, the main text, and its relation to the diagram portion of Figure 3.

Figure 4. Footnote 58 from Berge et al. (1984: 84)[Reproduced by permission]

The abstract graph portion of Figure 3 is itself a visual metaphor, showing a relation between position and temperature as if it were a shape in space, something that might be gestured and understood as a relationship among possible positions or motions. It brings to problems of continuous co-variation in all sorts of phenomena our intuitions and perceptual pattern-recognition capacities for spatial-motor phenomena. Those intuitions are supplemented by the mathematization of the abstract graph, itself an extension of an original, textualizable description of numerical data to a pure mathematical abstraction of pattern. Mathematics is more powerful than visualization, even though it is less intuitive, because it can represent patterns that cannot be visualized, and allow them to be compared, manipulated, combined, etc.

The abstract graph in this Figure is not mathematized in the text because it is much too simple. Any competent reader could do the mathematization immediately, but does not need to because this pattern is so familiar that it also has a verbal textualization, e.g. "The temperature rises linearly with depth in the fluid from To to To+dT." Or any one of many thematically equivalent textualizations, such as: "The temperature is directly proportional to depth in the fluid," or "The temperature decreases linearly with height in the fluid," or "There is a linear vertical temperature gradient in the fluid, with the bottom at higher temperature." Any competent reader could draw a substantially identical graph, or write the equivalent equation, from any of these verbal statements.

The abstract thematic formation which unites all these representations is semiotically multimodal, as it is in principle for all scientific concepts and their relationships.

Figure, Caption, and Text

Consider, as a final example, Figure 5, which reproduces Figure 1 of an article from Nature on the spontaneous movement of a biological molecule, kinesin (Svoboda et al. 1993). Careful inspection will show a subdivision, visually and by labels, into Figures 1a, 1b, and 1c (as they are labeled in the original), of which the latter two are of the data-graph genre, and Figure 1a is an extraordinarily complex visual integration of abstract diagrams and relatively abstract drawings of several sorts, as well as one semi-abstract graph. Important as visual codes, especially presentational and organizational ones (orientational ones, too) are to the interpretation of Figure 1a, it could not be readily interpreted without substantial reliance on thematic formations formulated in the semantics of natural language and indexed by the verbal labels in the Figure. For many of these, recourse to at least the caption text, and perhaps to the Main Text, is necessary, as well as prodigious intertextual connections to complete the formations and unpack their thematic condensations (Lemke 1988a): all that is left unsaid for the expert reader.

Figure 5. Figure 1 from Svoboda et al. (1993: 723)[Reproduced from Nature, by permission]

Quite apart from the complexities of the implicit mathematizations here, there are some very simple questions about how scientific texts are to be read so as to canonically combine Figures with Captions with Main Text. When we read the Main Text, there are references outward to the Figures, such as "(see Figure V.3)". We need to use verbal thematic formations at that point to construct meaning connections between the Main

Text from which we have come and what we find in the Figure. Those connections are most readily made either through the Labels in the Figure or through the intermediary of the Caption text. Captions as extensive as this one (equivalent to almost a full page of Main Text) are somewhat unusual; it contains a great deal of very important information necessary for the interpretation of the Main Text, which certainly has major thematic gaps without it. In fact the visual Figure itself is probably necessary for an interpretation of the Main Text adequate to the purposes of professional readers, and this is not at all unusual. Visual figures in scientific text, and mathematical expressions also, are generally not redundant with verbal Main Text information. They do not simply "illustrate" the verbal text, they add important or necessary information, they complement the Main Text, and in many cases they complete it.

I want to consider here only a few points concerning the ways in which visual semiotics contribute to and interact with verbal semiotics in constructing presentational, orientational, and organizational aspects of the meaning of this multi-semiotic text. Accordingly, I will begin from the Figure rather than from the Main Text, and refer first to text from its Caption and only then to Main Text. My comments are based on an extensive analysis of the visual-verbal relations among these elements.

Since we have concentrated so far more on presentational and organizational aspects of meaning in these scientific texts, let me here begin with orientational aspects. Orientational meaning includes (1) the stance or attitude a text seems to take to its own presentational content (e.g. warrantability, importance, desirability, usuality); (2) the stance it takes to prospective readers/viewers (users; e.g. solidary or antagonistic, deferential or condescending), including their anticipated attitude to the presentational content; and (3) the orientation of the text to other possible attitudes and viewpoints in the intertextual community (e.g. opposed, allied, complementary). Orientations to users can be analyzed by considering what the text seems to construct as its ideal prospective user, e.g. what the user is expected to know or be able to construe, what the user’s expectations and attitudes are likely to be. Scientific research articles generally construct users as highly knowledgeable in relevant technical matters, and mildly critical or skeptical, especially of warrantability claims, but also of importance claims. It is generally assumed that text-voice and user-stance share the same values as to what is desirable in scientific research.

The orientational attitudes to presentational content which are most salient in this genre are evaluations of warrantability, usuality or surprisingness, and importance. Desirability is generally backgrounded since it is assumed not to be in contention between text-voice and user-stance. It does however appear to mark this presumptive solidarity (see also Lemke 1988a). Consider some of the ways in which importance is constructed visually in Svoboda et al.’s Figure 1.

Importance is generally signalled by various conventions of visual prominence (see analyses by Kress & van Leeuwen 1990, O’Toole 1994, Arnheim 1956, Tufte 1983), one of which is relative size.

 

Thus among the graphs in Figure 1, the inset graph in 1b is much smaller in scale than the other graphs and is thus marked as of less importance, particularly relative to the larger graph in 1b.

 

 

 

 

In the Caption Text for 1b, there is a title, followed by two sentences (27 words) on the larger graph, and then one (10 words) on the inset graph. In the Main Text, a paragraph begins with reference to the larger graph (one sentence), and then a sentence on the inset graph follows. To understand why text and users would evaluate the inset graph as of lesser importance requires intertextual reference to the canons of experimental science regarding the best measures of instrument sensitivity and the most honest and revealing ways to display data on this important issue (important for the warrant of experimental claims made by using the instrument).

In fact, rhetorically, a major claim of this article is not just for its factual findings, but for the "advantages" (p.722, a rare explicit claim of Desirability) of its instrumental design over those in use by other researchers. This in turn accounts in part for the visual prominence of the whole of Figure 1, which is quite visually striking as a composition compared to typical diagrams of experimental apparatus, as well as for the prominence of 2b, which has a high visual density (compare 2c) and such additional highlights as the arrow pointing to the peak of the graph.

When reading 2b, the two most important visual features, relative to the claim for sensitivity of the instrument, are the downward trend of the graph and the fact that it lies mostly beneath a dotted line (visually foregrounding a fact that would be evident without this feature) representing a 10:1 signal-to-noise ratio. The inset graph not only provides subsidiary information, it also presents it in a way that makes a strong visually intuitive case for the irrelevance of noise in the signal, even though canonically the main graph is considered the more reliable measure of this effect. The inset is "icing on the cake", a sort of intuitive backup to the more technical evidence of the main graph it accompanies.

It is interesting that while the expert reader can get far more from the main graph, its implications for the issue of sensitivity are more indirect and must be mediated by considerable verbal discourse and reference to mathematical operations and relationships (which are provided in the caption text, in its subsidiary "Methods" section). On the other hand, the inset graph can be read topologically and intuitively as

showing the relative size of signal and noise (the large wave of ups and downs is the signal, the lesser wiggles are minor distortions of it, the noise). The one shows what the other proves. A scientific understanding integrates both these modes of warrant.

The most visually prominent feature of the entire Figure is surely the abstract drawing I will call the sphere-and-rod element (in 1a), including the shaded circles and their plane just below. We need in fact some reference to organizational visual semiotics here just to determine what constitutes a likely thematic unit in this part of the page. Proximity of elements, the enclosure effect of the parallelogram, and the specifically organizational visual device of the "enlargement" or "detail" lines diverging at an angle from the specimen stage (to the left and below) contribute. The diverging lines subtend in their angle the visual unit we are describing. They also call for us to perform a very complex visual-motor operation, in imagination, to co-orient the two different visual spaces (that of the microscope stage and that of the sphere-and-rod) in a single coherent space of representation. This

is a practiced operation for expert users.

The sphere-and-rod region is visually prominent on the page in part because it contains the only 3-dimensional representation, as well as the largest areal concentration of ink (especially the large dark sphere), and the boldest surface and contrast (black-white) effects. The prominent visual vectors (cf. Arnheim 1956) of the arrow, the parallel line of the rod’s axis, and the near-parallel sides of the parallelogram all contribute to the effect. Why is this visual complex so important? Presentationally, we are seeing here the heart of the clever design of this experiment: a silica bead (the sphere) reflects laser light (cf. the highlight on the sphere, the "shadows" in the plane) as it is pulled (the arrow) along a microtubule (the rod) by a kinesin molecule (small black blob connecting rod and sphere). The experiment is about the properties of the kinesin molecule, and the main claim is about how this molecule moves along a microtubule, but it is not the biological molecule that is visually prominent here (you might mistake it for a slip of the draughtsman pen), it is the "optical trapping interferometer", the clever experimental design to study the molecule indirectly by observing the bead.

As in all semiotic constructions, the visual qualities of an element mean in relation to those of other elements, especially those in the same presentation. All visual presentational forms in Figure 1 are highly abstract. The most recognizable is the schematic drawing of a computer, which is reduced in relative importance despite (or because of) its likely visual interest, by being small in scale compared to the sphere-and-rod drawing, and rendered only 2-dimensionally. Abstract as these drawings are, they still show what counts in our culture as pictorial representation. By contrast, the "optical bench" representation (elements such as lenses, stacked vertically and connected by dashed line, to the right of the long gray vertical dividing bar at left) and its continuation at the top into an electronic circuit schematic ("photodetector"

"amplifier") are purely abstract diagrams. They too could have been presented in more realistic renderings, or even in a photograph.

At the far left we find an echo of the optical bench elements marking off the stages of transformation of the polarization of the laser light as it passes through the various lenses and prisms. The arrows and arrows-and-ellipses are conventional symbolic representations of polarization, descended from actual graphs. The lens or photodetector symbols follow arbitrary conventions, one residually iconic. The algebraic symbols of

the Volts formula (boxed, top center) descend from words and sentences of natural language, represented here by the orthography of the near-by word-labels. We have here a large swath of the spectrum from visual-pictorial representations to verbal-semantic ones. They have been visually integrated (organizationally) and modulated for relative importance (orientationally).

Note that the organizational aspects of this Figure, what we need to know in order to construe its wholes and parts, its sequences and internal relationships, depend not just on visual information and conventions. Much of the interpretation depends on thematic (presentational) information from the verbal text, or verbal intertexts, in some cases mediated by

mathematical operations. This is also true in reverse. Figure 1c provides a view of data on the basis of which the weak forces with which the beads were "trapped" by the beams of laser light could be calculated. The relevant Caption Text reads " ... [the bead’s] brownian motion [was] recorded, from which a power spectrum was computed. The corner frequency provides the ratio ... [needed for the calculation]." The corresponding Main Text has: " ...[bead] dynamics corresponds to brownian motion ..., which has a lorentzian power spectrum. Experimental spectra are well fitted by lorentzians, and the corner frequency provides the ratio ... [needed]."

Inspection of Figure 1c, even without a knowledge of the mathematics of the lorentzian function or its canonical graphical forms, indicates visually what must be meant by the "corner frequency" and lets us interpret the legend on the graph as telling us that 725 Hz is the exact frequency at which we can see the graph of the power spectrum suddenly bending, as if around a corner. On the one hand we could interpret "corner frequency" merely as a quirky name for an algebraic parameter in an equation (for the lorentzian function), but the expert user here is expected to also interpret it visually, and perhaps physically, in terms of the shape of the graph, and in terms of the randomized motion to which it corresponds. Some users will rely more on one realization of the underlying multimodal thematics (verbal, mathematical, visual, physical), others on another, and the same user may shift and combine these as the situation requires.

 

Making Meaning with Multiple Semiotics

I will not provide the details here for lack of space, but it is quite possible, working from the Caption Text (and, with fewer points of direct correspondence, from the Main Text) to establish item by item correspondences between verbal elements and structures and visual ones in the Figure, particularly for the description of the passage of light (and so of information about the bead and kinesin molecule) in its various states of polarization along the optical bench to the electronic analyzer. This shows particularly well how the user must integrate visual and verbal realizations of objects, concepts, relations, and processes in the joint interpretation of text and figure. On the basis of this and similar analyses, it seems most reasonable to assume a joint visual-verbal thematic formation as the basis for interpretation, integration, and intertextual (and interpictorial) contextualization in making sense with multimedia texts of this kind. Thus, the "concept" of polarization, for example, must have not only a verbal semantic component, based in its canonical semantic relations to its usual collocates and contrast items (cf. Lemke 1983, 1990), but it must also have a visual-presentational component with similar relational meaning, as well as a mathematical component, and a technical-operational (sensorimotor) one, in order to be canonically useful in making meaning with a multimedia scientific article such as this.

It is sometimes argued that the various representations of a "concept" are entirely "redundant" with one another, that they can be placed in one-to-one correspondence, so that meanings that can be made in one semiotic modality can be equally well made in the others. This is not the way scientific communication appears to work: meanings are made by the joint codeployment of two or more semiotic modalities, and such codeployment of resources is needed for canonical interpretation. In my opinion, semiotic modalities (e.g. language, depiction) are essentially incommensurable: no verbal text can construct the same meaning as a picture, no mathematical graph carries the same meaning as an equation, no verbal description makes the same sense as an action performed. As with all abstractions, we learn how to construct relations and connections among these different semiotic acts and their traces; we learn how to count them as the same for some restricted purposes. But insofar as their meanings are not determined entirely by these correspondences with one another, but also by their relations to other forms within their own semiotic (picture to picture, and word to word; cf. Saussure’s valeur principle), even what they mean within these correspondences adds specificity of meaning in each semiotic modality beyond the common meaning shared across modalities.

"Concepts" as they are used in scientific communication, and in scientific work generally, are not defined by the common denominator of their representations, but by the sum, the union of meanings implied by all these representations. Nothing is really being "re-presented" here; there is no separate entity, no pure mental idea, apart from the meanings made with the specific material systems of semiotic resources our culture provides us with. What we call the abstract concept is only a shorthand for a multimodal semiotic construction, a simultaneous and multiply articulated cluster of interdependent practices. At the level of specificity that matters here, the different semiotic constructions that together and in relation to one another constitute "the concept" have nothing in common; there is no common denominator, and certainly no higher Platonic idea of which they are each pale shadows. It is in the nature of scientific concepts that they are semiotically multimodal in this sense, and this may well be true in other systems of semiotic practices as well.

When we investigate how meaning is made, we can no longer assume that actual social meanings, materially made, consist only in the verbal-semantic and linguistic contextualizations (paradigmatic, syntagmatic, intertextual) by which we have previously defined them. We must now consider that meaning-in-use organizes, orients, and presents, directly or implicitly, through the resources of multiple semiotic systems. What the logocentrism of a few prominent genres of purely verbal text (e.g. unillustrated novels, academic articles in philosophy and the humanities) has distracted us from, perhaps the pervasive multimodal disposition of scientific and technical text can remind us of. As computer technologies make multimedia genres more convenient and accessible for all purposes, it will become increasingly important to understand how the resources of different semiotic systems have been and can be combined.

What the careful study of linguistic semiotics has taught us about semiotics in general must now be extended to other semiotic systems and modified appropriately to their specificity. Language, for all its power, has had relatively little to tell us about topological, as opposed to typological, modes of meaning-making. We understand far too little about quantitative and mathematical meaning, about affective and bodily meaning, about dynamic and emergent meaning, about all the dimensions of material interaction and social being that are not well represented in categorial terms.

Language, and typological modes of semiosis generally, have evolved to work in partnership with other, often more topologically grounded, semiotic systems. It is my hope that by studying scientific and computer hypermedia genres, we can explore the powerful mysteries of that partnership in order to better understand just how we make sense of our world and ourselves.

REFERENCES

Arnheim, R. (1956). Art and Visual Perception. London: Faber.

Bakhtin, Mikhail Mikhailovich. (1935). Discourse in the novel. In M. Holquist, Ed. The Dialogic Imagination (1981). Austin TX: University of Texas Press.

Bateson, G. (1972). Steps to an Ecology of Mind. New York: Ballantine.

Bazerman, C. (1988). Shaping Written Knowledge. Madison, Wis.: University of Wisconsin Press.

Berge, P., Pomeau, Y., & Vidal, C. (1984). Order Within Chaos: Towards a Deterministic Approach to Turbulence. New York: Wiley.

Bernoulli, Johan. (1734). De motu musculorum. Bound in Giovanni Alfonso Borelli, De motu animalium. Neapoli [Naples]: Felix Mosca. [Item 105 in the Cole Collection of the New York Academy of Medicine Library.]

Bourdieu, P. (1972). Outline of a Theory of Practice. Cambridge: Cambridge University Press.

--- (1990). The Logic of Practice. Stanford CA: Stanford University Press.

Dyson, A.H. (1991). Toward a reconceptualization of written language development. Linguistics and Education, 3, 139-162.

Gregory, M. (1967). Aspects of varieties differentiation. Journal of Linguistics 3: 177-(198.

Halliday, M.A.K. (1975). Learning How to Mean. London: Edward Arnold.

--- (1978). Language as Social Semiotic. London: Edward Arnold.

--- (1982). The de-automatization of grammar. In J. Anderson, Ed., Language Form and Linguistic Variation, pp.129-159. Amsterdam: John Benjamins.

--- (1985). An Introduction to Functional Grammar. London: Edward Arnold.

--- (1991). Towards probabilistic interpretations. In E. Ventola, Ed., Recent Systemic and Other Functional Views on Language. (pp. 39-62). Berlin-New York: Mouton/deGruyter.

--- (1992). "How do you mean?" In M. Davies and L. Ravelli, (Eds.), Advances in Systemic Linguistics: Theory and Practice. pp. 20-36. London: Pinter.

Halliday, M.A.K. and R. Hasan. (1976). Cohesion in English. London: Longman.

--- (1989). Language, Context, and Text. London: Oxford University Press.

Hasan, R. (1995). The conception of context in text. In M. Gregory and P. Fries, Eds., Discourse in Society: Functional Perspectives. Norwood, NJ: Ablex Publishing.

--- (1984). "Coherence and Cohesive Harmony." Understanding Reading Comprehension. Ed. J. Flood. Newark, DE: International Reading Association.

Hicks, D. & Kanevsky, R. (1992). Ninja Turtles and other superheroes: A case study of one literacy learner. Linguistics and Education, 4, 59-106.

Hodge, R. & Kress, G. (1988). Social Semiotics. Ithaca: Cornell University Press.

Hymes, D. (1981). In Vain I Tried to Tell You. Philadelphia: University of Pennsylvania Press.

Kendon, A. (1990). Conducting interaction. London: Cambridge University Press.

Kress, G. & van Leeuwen, T. (1990). Reading images. Geelong, Vic: Deakin University Press.

Kress, G. & Threadgold, T. (1988). Towards a social theory of genre. Southern Review, 21, 215-243.

Latour, B. (1987). Science in Action. Cambridge, MA: Harvard University Press.

Lemke, J.L. (1983). Thematic analysis: Systems, structures, and strategies. Semiotic Inquiry 3(2): 159-187.

--- (1984). Semiotics and Education. Monograph in Toronto Semiotic Circle Monographs Series, Victoria University, Toronto.

--- (1985). "Ideology, intertextuality, and the notion of register" In J.D. Benson and W.S. Greaves, Eds., Systemic Perspectives on Discourse (pp.275-294). Norwoord, NJ: Ablex.

--- (1987). "Strategic Deployment of Speech and Action: A Sociosemiotic Analysis" In J.Evans and J.Deely, Eds., Semiotics (1983: Proceedings of the Semiotic Society of America `Snowbird’ Conference (pp.67-79). University Press of America.

--- (1988a). "Discourses in conflict: Heteroglossia and text semantics." In J. D. Benson and W. S. Greaves, Eds. Systemic Functional Approaches to Discourse (pp. 29-50). Norwood, NJ: Ablex Publishing.

--- (1988b). "Text structure and text semantics" In R. Veltman and E. Steiner, Eds. Pragmatics, Discourse, and Text (pp. 158-170). London: Pinter.

--- (1989a). "Social semiotics: A new model for literacy education" In D. Bloome, Ed. Classrooms and Literacy (pp. 289-309). Norwood, NJ: Ablex Publishing.

--- (1989b). "Semantics and Social Values." WORD 40(1-2): 37-50.

--- (1990). Talking Science: Language, Learning, and Values. Norwood, NJ: Ablex Publishing.

--- (1991). "Text Production and Dynamic Text Semantics." In E. Ventola, Ed. Functional and Systemic Linguistics: Approaches and Uses. [pp. 23-38]. Berlin: Mouton/deGruyter (Trends in Linguistics: Studies and Monographs 55).

--- (1992). "Interpersonal Meaning in Discourse: Value Orientations." In M. Davies and L. Ravelli, Eds. Advances in Systemic Linguistics: Recent Theory and Practice. [pp.82-104]. London: Pinter.

--- (1993a). "Making Meaning with Language and Other Semiotic Systems" Paper presented at International Congress of Systemic and Functional Lingusitics, Victoria BC, Canada.

--- (1993b). "Discourse, Dynamics, and Social Change." Cultural Dynamics 6(1): 243-275. [Special issue, Language as Cultural Dynamic, M.A.K. Halliday, Issue Editor].

--- (1993c). "Intertextuality and Educational Research." Linguistics and Education 4(3-4): 257-268.

--- (1993d). "Hypermedia and Higher Education." Interpersonal Communication and Technology (IPCT-L@GUVM), April 1993. [To be printed in T.M. Harrison and T.D.

Stephen, Eds. Computer Networking and Scholarship in the 21 st Century University . Albany: SUNY Press.]

--- (1994). "Multiplying Meaning: Literacy in a Multimedia World." Paper presented at the National Reading Conference, Charleston SC (December (1993). Arlington VA: ERIC Documents Service (ED 365 940).

--- (1995a). Textual Politics: Discourse and Social Theory. London: Taylor and Francis.

--- (1995b). "Intertextuality and text semantics." In M. Gregory and P. Fries, Eds. Discourse in Society: Functional Perspectives Norwood, NJ: Ablex Publishing.

--- (1996). "Self-Organization and Psychological Theory." Review of L.B. Smith & E. Thelen, Eds., A Dynamic Systems Approach to Development. Theory & Psychology, 6(2): 352-356.

--- (in press, a). "Semantic Topography and Textual Meaning" in R. Stainton & J. DeVilliers, Eds., Communication and Linguistics. Toronto: Glendon College, York University. [Festschrift for Professor Michael Gregory]

--- (in press, b). "Typology, Topology, Topography: Genre Semantics" to appear in Network (Sydney: International Systemic Linguistics Assn.).

Mann, William and Thompson, Sandra. (1986). "Relational propositions in discourse." Discourse Processes 9(1): 57-90.

Martikainen, P.J., Nykanen, H., Crill, P. & Silvola, J. (1993). "Effects of a lowered water table on nitrous oxide fluxes from northern peatlands." Nature 366: 51-53.

Martin, J.R. (1991). Intrinsic functionality: Implications for contextual theory. Social Semiotics 1(1): 99-162.

--- (1992). English Text. Philadelphia: John Benjamins.

Matthiessen, C.M.I.M. (1992). "Interpreting the textual metafunction." In M. Davies and L. Ravelli, Eds. Advances in Systemic Linguistics: Recent Theory and Practice. [pp.37-81]. London: Pinter.

Ochs, E., Gonzales, P. & Jacoby, S. (in press, a). "When I come down I’m in the domain state ...: Grammar and graphic representation in the interpretive activity of physicists." To appear in E.Ochs. E. Schegloff, & S, Thompson (Eds.), Interaction and Grammar.

--- (in press, b) "Interpretive journeys: How physicists talk and travel through graphic space." To appear in Configurations 2(1).

O’Toole, M. (1990). "A systemic-functional semiotics of art." Semiotica 82: 185-209.

--- (1994). The Language of Displayed Art. London: Leicester University Press.

Peirce, C.S. (1955). "Logic as Semiotic: The Theory of Signs" in J. Buchler, Ed., Philosophical Writings of Peirce. New York: Dover.

Percival, I.C. (1987). "Chaos in Hamiltonian systems" in Berry, M.V., Percival, I.C. & Weiss, N.O., Eds., Dynamical Chaos. Princeton NJ: Princeton University Press. pp.131-143.

Scheflen, A.E. (1975). Models and epistemologies in the study of interaction. In A. Kendon, R. Harris, & M. Key, (Eds.), Organization of behavior in face-to-face interaction. The Hague: Mouton.

Svoboda, K., Schmidt, C.F., Schnapp, B.J., & Block, S.M. (1993). Direct observation of kinesin stepping by optical trapping interferometry. Nature 365(6448), 721-727.

Thibault, P. (1991). Social semiotics as praxis. Minneapolis: University of Minnesota Press.

--- (in press). "Interpersonal Meaning and the Global Modal Program" To appear in M. Gregory & M. Cummings, Eds. Relations and Functions Within and Around Language. Norwood, NJ: Ablex Publishing.

Thom, R. (1975). Structural Stability and Morphogenesis. Reading, MA: W.A. Benjamin.

Tufte, E. (1983). The Visual Display of Quantitative Information. Cheshire, Connecticut: Graphics Press.

Wilden, A. (1980). System and Structure. London: Tavistock.

SEMIOTICS AND THE DECONSTRUCTION OF CONCEPTUAL LEARNING

Originally published in J. Soc. for Accelerative Learning and Teaching, 1994

J. L. LEMKE

What is Postmodernism, and Why is it Saying all these Terrible Things?

Postmodernism is a loose alliance of intellectual perspectives which collectively pose a challenging critique of the most basic assumptions of the modern educational enterprise. What are these perspectives? What do they have to say that can open new doors for educational research and practice? I would first like to sketch a personal view of postmodernist discourse, and then focus on its challenges to the foundations of most modern views of abstract conceptual learning. Going beyond critique, I would like to sketch some alternative, postmodern possibilities.

Modernism, like any intellectual movement, will ultimately be defined from the viewpoint of its successors. Postmodernism is beginning this process by offering a critique, a reaction against some intellectual trends that perhaps began as early as the

Renaissance, but certainly became well established by the later 19th and early 20th centuries. This is the foundational period for the discipline of psychology and for educational theory and most of the social sciences, as well as for modern literary and fine arts criticism. Postmodernism is primarily, in its origins, a philosophical critique of the assumptions built into these disciplines in their formative years. The critique is gradually being extended today into the natural sciences, the inner fortress of modernist assumptions about knowledge and reality.

From the postmodern point-of-view, modernism is defined by its belief in objective knowledge, or at least in the possibility of objective knowledge, and by its assumption that such knowledge refers directly to an objective reality which would appear in the same way to any observer. A further characteristic modernist assumption is that knowledge is a product of the activity of the individual mind, fashioning its ideas or mental schemas to correspond with this objective reality.

Postmodernism, on the other hand, argues that what we call knowledge is a special kind of story, a text or discourse that puts together words and images in ways that seem pleasing or useful to a particular culture, or even just to some relatively powerful members of that culture. It denies that we can have objective knowledge, because what we call knowledge has to be made with the linguistic and other meaning-making resources of a particular culture, and different cultures can see the world in very different ways, all of which "work" in their own terms. It argues that the belief that one particular culture's view of the world is also universally "true" was a politically convenient assumption for Europe's imperial ambitions of the past, but has no firm intellectual basis.

Many postmodernists go further and point out that just as Europeans temporarily imposed their view on other cultures by force, so within European cultures, the upper social classes, and particularly middle-aged, masculinized males have dominated the natural and social sciences (as well as politics and business), and so this would-be-universal worldview is even more narrowly just the viewpoint of one dominant social caste or subculture.

Postmodernism traces its roots to the movements of structuralism and its counter-reaction, post-structuralism, mainly in the the French-speaking intellectual community of the 1960s and 1970s. Structuralism is either the last stage of modernism or the immediate precursor of postmodernism. The great spokesman for structuralism in this period was the anthropologist Claude Levi-Strauss (e.g. 1963), who argued that the patterns of human culture, from village architecture to premodern myths, had the subtle regularities of mathematical structures (not quantitative ones, but combinatorial and symmetry patterns as in algebraic group theory). He was joined in this by Jean Piaget, who saw in the developmental movement of childen toward modern concepts of space, time, matter, and quantity a construction by the child's mind of just these same sorts of regularities in the operations it performed to make sense of the world (e.g. 1970, 1971).

The structuralists were often inspired by the successes of modern linguistics, first in using meaning to analyze the seemingly mathematical regularities of the sound systems of language (e.g. Jakobson 1956, 1962), and later in Chomsky's (1957, 1965) use of transformations (the heart of quasi-mathematical structuralism) to illuminate regularities

in syntactic rules. Language, culture, and thought were all to be brought at last into the modern fold of mathematically regular sciences.

By codifying the enterprise of modernism, the structuralists made it a more precise target for its critics. The critique can trace its roots, a bit mythologically, back into the Enlightenment (e.g. Giambattista Vico 1774/1968), and with some justification as far back as Friedrich Nietzsche (e.g. 1967, 1989). Those who have been posthumously enrolled in the pedigree of postmodernism include the philosophical phenomenologists (e.g. Edmund Husserl, Maurice Merleau-Ponty, Jose Ortega y Gassett) and the great apostate of modernism, Ludwig Wittgenstein (1949/1958). But the most influential, and first truly post-modern critiques were those by Michel Foucault (esp. 1969/72) and Jacques Derrida (esp. 1967/1976).

The common denominator in all these critiques of (except perhaps for those of the phenomenologists, who were recruited rather late) is the analysis of discourse. It is the focus on discourse, and language, that unites structuralists and post-structuralists, but while the structuralists sought the regularities of language, the post-structuralists pointed out how language-in-use, the discourses of philosophers and poets (and, yes, of psychologists and physicists, too), refuse to be contained within the patterns of order we try to impose on them. To the extent that our discourses are our tools, if our tools are -- the modernist would say, in horror, unreliable; the postmodernist, in exaltation, self-organizing -- then the scientific ideals of the structuralists, and of modernism generally, are a chimera, an immature self-delusion we have to grow out of.

Foucault said, in effect, that it was chimerical to imagine that historians could reconstruct a real past; historical discourse is a discourse of the present, serving present ends, making sense for us today out of the archeological traces of past human activity. Foucault undertook to write histories of discourses, including in that term not just language, but all the things people do to make sense of their worlds. He showed, to many people's pleasure, that the very objects of modernist scientific investigation, the very notions of self, nation, language, mind, sex, crime, normality were themselves specific historical constructions, the products, not the objects of human discourse and inquiry. He refocussed attention from the so-called `phenomena' science sought to investigate, to how science (read philosophy, psychology, law, common culture) produced meaningful phenomena through its discourses.

The essential step away from modernism was the new focus on meaning. How does a text mean? How does a graph or diagram tell us something? How do marks on paper (or lighted pixels on a screen) convey to us a complex conceptual meaning? A landscape is a text for a geologist. An apparatus and its readouts is a text in the same sense for a physicist. The principles, and the problems of meaning, are the same. If the meaning of any text depends on how we interpret it in relation to other texts, how can either data or explanation be fixed and stable in its own meaning, much less the basis for objective knowledge of an objective world? Why should we believe that scientific texts are not subject to the same principles of interpretation as literary texts? Why should we believe that primary data can be read from the book of nature without the same problems, or arbitrary conventions, of interpretation that beset the reading of any other sort of "book"? Why should we believe that the practical effects of technologies are legitimate warrant for the objective truth of theories, when the only links between them are yet more texts, more discourses?

Derrida was more playful in his critique, and more direct. Foucault the historian had shown, again and again, exactly how the disciplines created their objects through discourse, and created them very differently in different historical periods. Derrida was a philosopher and sought, in his disruptive readings of classic texts, to show how imperfectly any discourse makes its objects and the world they are supposed to inhabit. Whether a literary, philosophical, or scientific text, Derrida "deconstructed" their constructions of "real objects" of study or narration. He took hold of the key structuralist notion of transformation (A -> B), and showed how unstable any discursive construction of "difference" (A - B) must be. With a rather Zen-like sensibility, he focussed on the absences (B is not A) that all presences presuppose, and on the gaps that must be made to separate things we wish to construe as different. He attacked not just positivism (the naive epistemology of turn-of-the-century science, and still of most contemporary curricula), but "positivities", the notion that things are to be defined by what they are, when in fact discourse can only define something as what it is not. (All categories are based on sets of contrasting alternatives; to be of type Y means not being X or Z.)

Derrida is not widely popular among English-speaking academics -- partly, I think, because they have no patience with the essentially literary genres in which French-speaking intellectuals write philosophy. His deconstructionism, primarily a movement in literary criticism and similar disciplines, is very threatening when applied to the social sciences. It undercuts the possibility of their being positive sciences at all, of their pretensions to "objective knowledge" about the social world. But then deconstruction raises the same epistemological problems for the traditional sciences as well, without negating the usefulness of their technologies. (Once the link between theory and practice is weakened, theory cannot turn to practice for its "proof", but then by the same token practice can go on its merry way whether the metaphysical claims of theory are justified or not.)

There are two other major currents in postmodernism: phenomenology and semiotics. Phenomenology (for major figures, see above) basically chides the scientific view of human life with being too narrow, too mechanical, too little focussed on how people live and make their lives meaningful, too obsessed instead with the artificial objects made by their discourses: kinship systems, cognitive schemata, class struggles. Feminism has found phenomenological perspectives congenial for efforts to construct new discourses of the world rooted in women's experience; intellectual feminists are attracted to postmodernism because it makes it much easier to see how what men define as "the world", "the problems", "the disciplines", are just that, what some men's discourses have defined them to be. (See for example Harding 1986, Haraway 1991, Nicholson 1990.)

The phenomenological perspective does not need to be limited to conceptualizing how the world looks different to men and women; it can be used to examine how it looks different to the young and the middle-aged, to the novice and the expert, the student and the teacher, the ghetto child and the comfortable academic. We each construct our own lifeworlds, and even when we are in the same room, trying to talk to one another, we may still be worlds apart.

Semiotics, the last of the major currents within postmodernism that I will mention, is a generalization of linguistics. At its narrowest, it is merely a codification of the symbols

offered us by our culture, and the formal description of how those symbols are usually combined, be they words, gestures, graphics, foods or clothes. At its most general and most powerful, it is the analysis of how we deploy our cultural resources for making sense of the world: language, depiction, action. It is the systematic uptake of Foucault's challenge: to see how we make our meaning-reality with the symbolic tools available, and how our doing so leads to changes in those tools, our use of them, and the sense we make with them. Because semiotics is the branch of postmodernism I know best, and the branch that offers, I believe, the most highly developed alternative tools for educational research and practice, I want to describe it a little more, and turn its light on the problem of abstract conceptual learning.

Social Semiotics and the Construction of Meaning

Linguistics is important because language is, par excellence, the tool most of us use to make sense of nearly everything nearly all the time. Language is a resource for making meaning. Studying how people use language (i.e. our discourse, in the narrow, linguistic sense) to make the meanings of physics, or economics, or cognitive psychology, in research, in teaching, in writing, in dialogue with colleagues and students, can reveal how these disciplines construe general and special human experience into the categories and relations that characterize their unique disciplinary perspectives (some examples: McCloskey 1985; Bazerman 1988; Lemke 1988a, 1990a, 1990b; Halliday 1988; Halliday & Martin 1993). This is linguistics focussed not on form, but on function (e.g. Halliday 1985). It has more in common with the tradition of grammar-and-rhetoric than with modern formal syntax (though there is a gradual convergence). Linguistic discourse analysis is an applied semantics, a textual semantics, with formal syntax functioning as the legs that carry the dog, but not the tail that wags it.

Some studies of discourse(s) look more closely at the grammar-semantics connection (Halliday 1988, Halliday & Martin 1993); others look more at the semantics-rhetoric connection (Lemke 1988a, 1990a, 1990b), or at the rhetoric-genre connection (McCloskey 1985; Bazerman 1988). If I write a science textbook, my editors and readers expect me to follow certain conventions of format and organization: certain sorts of information are to be included, laid out in a certain way. These are the conventions of the textbook genre. In the course of a chapter I may want to persuade my readers to accept an argument, or I may want to raise some critical questions. These are functions of my rhetoric at that particular point, what I am trying to do for (or to) the reader. If I am going to be persuasive I need to follow the conventions of logic and argumentation that my community of readers probably accept and are used to. As I write my argument, I have to express particular meanings about something. I have to choose my topic and what I want to say about it specifically. These are matters of the semantics of my text. They influence my choice of words and of conceptual relationships that I wish to express. In order to do so, in writing a particular clause of a particular sentence, I also need to bear in mind the rules of English grammar and its syntax.

Studies of discourse are inevitably embedded in studies of social and cultural conventions; they require a social linguistics more than a cognitive linguistics (though both may have their uses). It is not possible to adequately analyze how individuals make sense if you do not know what the typical discourse patterns, the typical sense-making practices, of their community are (their semantic, rhetorical, and genre conventions at

least; see for example Lemke 1989). It is particularly obvious that you cannot do so when you see that all discourse analysis is founded, explicitly, or implicitly, on the principle of general intertextuality (Lemke 1985): all meaning is made against the background of other meanings already made and shared in a community.

When I write the textbook chapter, or read one like it, I make sense of it by comparing it, consciously or unconsciously, with other similar texts I have read before. I recognize that it is a textbook and not a research treatise by its genre conventions, so I expect some things from it and not others. I make sense of its content largely because I can fill in its unstated assumptions from other works I have read on the same topic. I make sense of its patterns of argumentation because they are familiar to me from elsewhere. I interpret its use of technical vocabulary based on the specialized meanings I know these words have in this specialty field. And so on.

The totemic grandfather of social linguistics and intertextual discourse analysis of this sort is Mikhail Bakhtin (1929, 1981, 1986), a figure who ranks today with Foucault (and perhaps outranks Derrida). Like Foucault, Bakhtin takes a larger view of discourse than simply language in its narrowest sense. Discourse is a mode of action, almost synonymous with meaning-making itself. The units of discoure for both of them are units of meaning, or units of human activity that make meaning (utterances), and not linguistic units per se (clauses, sentences). Their perspective is semiotic. Semiosis is the process of making meaning by deploying the resources of social systems of signs in a community. While linguistic signs (words, clauses, texts) form such a semiotic resource system, so do many nonlinguistic, or only partly linguistic modes of human action. We can make meaning with dance, gesture, and movement; with pictures, diagrams, and typefaces; with songs, meals, and clothes. Most fundamentally, we make meaning with action. Linguistics made the first breakthroughs in the study of how we make meaning by deploying semiotic resources, but the general processes, it appears, apply to all meaning-making activity.

In the sciences, we do not just talk and write. We also act in many other ways that contribute to the special meaning constructions of our disciplines (e.g. Latour 1979, 1987) whether in the laboratory, in the field, in data collection and analysis, or in the economic, social, and political dimensions of the subcultures of our disciplinary institutions. All these actions function in the same way discourses do, to make possible the meanings of the discipline, to construct the objects we say we study. To say that objects or phenomena are social, discursive, actional, semiotic constructions is not to deny their materiality. It is to emphasize that what our theories, our discourses take hold of, are objects or phenomena as meanings, i.e. as we conceive of them, speak of them, measure them. This is not so far from the scientific doctrine of operationalism, that every physical object or quantity is defined by our procedures for observing or measuring it. Semiotics tries to tell us as much as possible about how it is possible to construct such meaning-objects.

I can read a book or I can weigh it on a scale. I can treat it as a symbol to be interpreted semiotically or as a material object to interact with physically. Semiotics is a discourse that tells me how the book's writing and diagrams and language mean something. Physics is a discourse that tells me how books and springs and scale-pointers interact physically, and which enables me to read the scale and interpret it as having a meaning

about the weight of the book in the context of a discourse about mass, weight, gravity, force, elasticity, etc.

Every symbol must also be a material object, and any material object is recognizable and interpretable only in relation to some system of semiotic categories. Every act of interpretation of a symbol is also a physical, physiological process as well as being a semiotic practice, an enactment of a cultural system of conventions for making meaning. The material and semiotic aspects of things and processes are complementary to one another. They describe two different systems of relationships that we can construct among objects and processes. One of these is the familiar system of material, physical, chemical, thermodynamic, ecological relations: webs of material interaction. The other is the semiotic system of relationships of meaning: similarity, difference, categorization, ordering, association, etc.

The most basic semiotic relationships are very abstract, but for many purposes one can think of them in terms of a few simple types. Paradigmatic relations are those that tell us what something might have been but isn't: a physics textbook is not a biology textbook, a textbook is not a novel, a book is not a magazine, etc. Paradigmatic relations define contrasting alternatives, meaningful differences within similarity. Syntagmatic relations tell us what parts make up some whole: words that form a single sentence, sentences that form a single paragraph, different volumes that make up a single encyclopedia. Intertextual, or indexical, relations tell us in a broader sense what goes with what: this book is relevant to interpreting that book, this situation or event is a relevant context for that one, etc. The most fundamental principle of semiotics is that meaning is possible only because not all possible combinations of things, events, contexts, are equally likely. The particular odds on various combinations describe the culture of our community: our expectations and our patterns of behavior, including how we interpret meanings and how we interact with our environment.

A postmodern semiotic constructivism such as I have just described (for a fuller account, see Lemke in press) is itself, of course, just another discourse. You can get used to it, use it, enjoy it. Or find another that works better for your purposes. Postmodernism reclaims for science, and philosophy, the intellectual freedom of art. It refuses the power moves of some discourse factions that insist that their discourses are the only possible ones because they are "true". Many different discourses "work". It is not even possible to say absolutely whether many of these discourses are "consistent" with each other or "incompatible". They can be construed as being either, and usually are.

Most people reared in modern positivist traditions, or their commonsense variants, find themselves viscerally upset by the idea of rejecting notions like objective truth and reality. For a long time we have been sold the belief that these notions are indispensable, not only for science, but for morality. Postmodernist sensibility regards these reactions as understandable, but a little childish. Many postmodernists are arrogant; high levels of self-confidence are necessary when disagreeing with the foundations of a large part of your own culture. But postmodernism, semiotics, or social constructivist epistemology, do not require total Faith, or even complete self-consistency (another chimera, read G”del 1962). You can continue to believe that there is an objective external reality out there somewhere, and that truth is the common quality of propositions that correctly describe it, so long as you do not use these

assumptions to try to gain power over your opponents in intellectual debate. Many postmodernists believe that such assumptions actually have no other practical function. Agnostics are heartily welcome in this un-church.

The postmodern, semiotic, constructivist view talks about meaning, not about truth. It talks about how discourses define phenomena, not about how phenomena are described by discourses. It always wants to know what people do that makes sense of what we ordinarily call an object or phenomenon. It situates meaning-making practices and the systems of semiotic resources deployed in those practices in the domain of the social, the cultural. Indeed, it sees social and cultural systems exactly as systems of such practices, systems of doings, and not systems of doers per se. The doer, the notion of a human individual, is as much a meaning-construction as anything else. If it is doings, i.e. social and cultural practices, that are fundamental, then as activities these practices consist of processes and participants defined in relation to the processes. Among the types of participant constructed in our culture are ones we call human individuals, but what a human is (an organism, a social individual, an actor or agent) is not necessarily the same from one type of activity to another. We learn how to conflate them, to make them all seem the same, and indeed how to think of ourselves as being constructions of this kind (cf. Lemke 1988b, in press). Human individuals cannot be taken for granted as the starting point of either social or cognitive theories.

As an example of this consider the question of whether and in exactly what sense a student is the same person in class and out of school, in math class and in English class, in small-group work and in whole-class instruction. Of course our culture provides ways to unify these differently behaving individuals, but it takes work to do so, semiotic work. We miss an important perspective on the student if we carry this presumption of unity and consistency too far. If we assume that the student has the same characteristics, the same ability, the same intelligence, even the same personality or interests, in all these different settings and situations, we may be overdoing it. And if we assign a grade to a student, instead of to an event or a performance, what does it really mean, if the student wasn't the same person in each different situation on which we are basing the single grade?

Semiotic Perspectives on Learning and Abstraction

The dominant theory of learning that guides educational practice in our society says that what people need to learn are "abstract concepts," which they can then apply to a wide variety of specific situations. Nearly everyone is convinced that conceptual learning is the most powerful form of learning, and the only problem is how to get more people to be able to successfully learn abstract concepts. The criterion for having learned an abstract concept is being able to apply it in new, unfamiliar situations. The way to teach abstract concepts is to demonstrate how they apply to several different situations until the student "catches on" or generalizes and "gets" the concept at an abstract level. The student will then be able to use the concept wherever it is relevant. -- Do you believe this fairy tale?

What happens in practice? A very, very small percentage of students seem to be "able" to learn abstract concepts in the sense described. If we accept fairly weak criteria of conceptual mastery, say the ability to apply the concept in situations not too different from the ones in which it was taught, but are rigorous about mastery at this level, a

reasonable estimator of success across the student population would be the numbers of students who get top grades in courses like algebra or physics. Even most of these students, however, or their counterparts in other disciplines, would not meet the standard of being able to apply the concept `wherever it is relevant'.

Applying the older theory, we could say that either there is something wrong with the students, or something wrong with the teaching (or testing) methods. But the evidence against this prescription for abstract conceptual learning for most students is so overwhelming that surely we ought to consider that there might be something fundamentally wrong with the theory? The closest that traditional educational psychology comes to this is the great "transfer of training" debate. The only conclusion I have been able to draw from all the data on both sides of this is that, in general, the more abstract the concept, and the more unfamiliar the application context or content, the less evidence there is for generalization.

A now long and distinguished tradition of dissenters (e.g. from Cole et al. 1971, Cole & Scribner 1974, to Lave 1988) have argued that higher reasoning processes are context- and content-sensitive, not context- and content-independent. People in this tradition have even wondered whether cognitive processes can be usefully described at all apart from specific social and cultural activities, or at least apart from relatively specific social and cultural strategies for action.

The strongest evidence for conceptual generalization seems to come either from low-level processes (e.g. perceptual shape recognition), where evolution may well have lent a helping hand, or from our persistent introspection which tells (a very, very few) of us that this is what we do in higher-level reasoning. But our introspection is tainted by the theory itself: we have ourselves internalized a common discourse of our academic, intellectual culture which shapes the meanings we give to our subjective experience. Perhaps neither we nor anybody else does any such thing as generalize an abstract concept. Perhaps there is no such thing as an abstract concept.

Put that way, we recognize that "abstract concepts" are not the sort of thing to which our culture assigns much of a "reality" status anyway. Do abstract concepts exist? We would probably say that, no, perhaps the phenomena they refer to, or describe, do, but the concepts themselves are just a shorthand way of talking about rather complex cognitive processes. The postmodern semiotician, of course, will see the very notion of "abstract concepts" as born of the discourse of cognitive psychology (and its predecessors), and regard what goes under that name as just a set of conventions for using particular linguistic forms (often in conjunction with nonlinguistic actions of various sorts). Some of us (myself included) question the usefulness of even the notion of cognitive conceptual processes as such, wondering what, at least in the present discussion, cognitive theories can say about human reasoning that linguistic discourse analysis does not describe more precisely and with greater economy of theoretical means (especially remembering that the data of this area of cognitive psychology are, by and large, contextualized verbal reports in the first place; see Thibault 1986; Lemke 1989, 1990a; and even Geertz 1983).

In education we often proceed to teach something by first `breaking it down into simple parts or steps' then teaching the steps and expecting students to be able to perform the whole. In the teaching of skills (famously, bicycle riding), it is well known that this is

not effective, but in the teaching of most academic intellectual skills it is still routinely adhered to. Constructivism points out that this sort of reductive analysis into parts is only possible as a post-hoc activity. It is only after we have mastered the whole that we can understand how it can be artificially divided into parts. The parts are not `natural', they are not there as a given prior reality. The parts are constructed by local conventions that depend on a prior facility with the whole. You cannot learn wholes through their parts. And the reason is that they don't have these parts! What you can do is learn how wholes are conventionally analyzed into parts, learning what precise kinds of part-whole relationships need to be constructed.

In academic education we also assume that students can learn abstract principles by induction from examples and by descriptions of abstract properties and relations. But just as skills do not necessarily have specific "parts" apart from how we choose to analyze them into these parts, so also an abstract principle is not necessarily visible in its "examples" until we learn how these examples are conventionally construed as instances of the same general principle. It is not necessarily true that the principle is "there" in the examples to be seen by anyone. Many students don't see these imaginary properties of examples even when teachers try their best to point them out.

In a classroom episode I analyzed a few years ago (Lemke 1990: 144-148), students could not "see" a wave moving on a long coiled spring in the way the teacher did, despite the teacher repeatedly demonstrating it right in front of them. They had to learn to "see" it in a new way, mediated by special technical distinctions named by specific terminology. They had to learn to use language, in conjunction with vision and motor action, to reinterpret experience in a new way, to "see" something that for the teacher was simply "there" in front of them.

For the purposes of learning and social behavior, we do not simply "see" photons registering on our retinas; we "see" meaningful patterns created by the higher centers of our brains according to the habits and conventions of our culture. The way in which these patterns are constructed is still somewhat mysterious in neurological terms (see Edelman 1992), but the social evidence for the process clearly shows the role of language and other systems of symbols. Learning to use a semantic distinction, such as that between "longitudinal" and "transverse" wave, or that between "motion of the medium" and "motion of the disturbance," as part of language is an integral part of learning to make and use the conceptual distinction.

We expect students to `catch on', to formulate abstract generalizations that will then apply to new and unfamiliar examples. We expect that they will `transfer' the abstract principle to new settings. But why? Mainly because our own cultural traditions, from Platonism to positivism, assume that the situations in which the principle applies really are "the same" in some respect that we can learn to recognize.

But for the postmodern constructivist there are no inherent similarities except the ones that a culture, a community constructs as meaningful, as significant against the background of an infinite number of possible categorizations, and constructs always, again, post hoc, i.e. after each instance is encountered. What our semiotic practices, such as the use of semantic distinctions coded in language, do is to enable us to fit instances into prior categories, or to create categories to encompass known sets of instances. We must invent a way to fit each new type of instance into an existing

category, and insofar as the category is defined by the practices that assign its members, we actually change the category (i.e. add new categorization practices) for each new type of member.

This can only be done post hoc. Only instances that we already know how to type as members of a category will accomodate `transfer'; genuinely non-trivial new instances cannot be automatically typed because they do not already have the categorially critical features -- those features must be constructed for them. They do not automatically `fit' the category; they must be fitted into it. That requires work, social work. It cannot be done by an individual, because it must be done by the conventions of a community. It is ultimately as much or more a social than a `cognitive' process.

As an example, consider one of the most widely generalized concepts of natural science, "energy". The history of science shows clearly that each new `form' of energy (sound, heat, light, etc.) had to be defined in just such a way that it could be assimilated to the existing concept of energy. In fact, as each new form of energy was added, the concept itself, insofar as its meaning can be described as the sum total of all its possible uses or operational definitions, changed. The concept also changed insofar as it is defined by the set of operations for applying it to various phenomena.

No one can be expected to `generalize' from a notion of kinetic energy to the concept of potential energy, or to be able to anticipate the proper definitions of various forms of nuclear energy from a knowledge of heat energy or electromagnetic energy. What do all the forms of `energy' have in common? not even how they are measured, not even their `operational definitions'. You have to learn to call them all "energy" and learn that what this means is that, properly defined or measured, they can all play the same role in particular calculations or arguments, all fill the same slot in particular theoretical discussions. Our community has developed conventions for fitting each variation on the theme of energy into a common semiotic pattern (i.e. a general set of statements, whether linguistic, mathematical or graphical). That development has been the work of centuries. It can be recapitulated, but it would be foolish to imagine that it can be anticipated by individuals because it is somehow `there' in nature to be seen. In fact, you have to be carefully taught how to look to see `it' (i.e. to construct it) or how to `show' it to (construct it for) others, or even how to argue that others should accept what you show as evidence for what you claim.

The implications for education of this reconceptualization of what abstract concepts are is radical and profound. In this model generalizations are culture-specific: of all the possible similarities of two events, two moving springs, two "systems with energy", our culture has historically opted to pay attention to just certain ones, which it has evolved methods (semiotic methods: linguistic, experimental, graphical, mathematical) for constructing. Even if you believe that the similarities are "there" in the sense that it just wouldn't work for practical purposes (whose purposes? how practical?) to construct ANY old similarity, there are still an awful lot of possible similarities that CAN be constructed between two events or two systems in nature. Learning our culture means learning which ones we do construct, how we construct them, and what good they are for our practical purposes.

But if these particular similarities are not "obvious" ones, if we have to learn how to "see" (i.e. construct) them, then the process of learning "an abstract principle" or "an

abstract concept" or generalization is really the process of learning how to construct specific sorts of similarities among specific classes of instances. In terms of classical logic this means that categories are learned "extensionally" by learning what their members are and why they are members, rather than "intensionally" by learning a set of features which are common to all members of the set. Those "same" features have to be constructed DIFFERENTLY for different members of the set. We probably do this by stages or degrees, first learning how to "see" some sorts of phenomena as "waves", then how to see other sorts as being similar, then how to see still other sorts, etc.

When we come to a new and unfamiliar class of phenomena, we can propose that these too are waves, but it is ultimately a matter of social consensus whether our proposal is adopted or not (for whatever reasons). If history has already made this decision for us, and the criteria of the culture for validating that decision leave no leeway for reconsideration, there is really no way to expect us, the students, to second-guess this history. We just have to be told. We have to be let in on one more set of specific procedures for how to make this new class of examples look like the others, how to construct similarities between it and other classes of examples.

Perhaps those of us, a small minority, who are positioned within our culture in such a way as to have acquired habits for guessing (or reasoning) that are most similar to those of the people who made these determinations historically (sc. upper-middle class, masculinized, middle-aged Northern European males), we are more likely to guess "right", i.e. to decide as our forbears did. This does not make us more intelligent or more able at abstract thinking. It only signals that we have been cut from the same cloth. (Such habits are largely unconscious, and the product of experiences in all aspects of our lives, not just school experiences. For a theory of them as embodied dispositions, see Bourdieu 1990; for gender and class differences in how children and students make meanings with language in interactions with mothers and with teachers see Hasan 1986, 1990, 1992 on semantic orientations.)

Another important implication of this reconsideration of the nature of abstract conceptual learning is that the value of studying something "similar" to our ultimate object of interest is called seriously into question. In academic learning we have acquired the habit of teaching by simulations and simulacra, rather than by giving our students first-hand experience of "the real thing".

In science education, for example, we expose students to science textbooks instead of, say, to scientific text; to science teachers instead of to scientists; to school laboratories instead of to scientific and technological workplaces. Simplified equipment, simplified procedures and processes, whether intellectual, conceptual, or manual will not suffice. They may have a function as adjuncts to learning, once students have already participated in the actual social practices being taught, but we cannot expect them to function, as they do now, as substitutes for such direct participation. It is only after we have learned how, say, science and technology operate in our communities in real laboratories and workplaces that we can intelligently participate in the construction of correspondences and similarities between what happens there and what happens in science classrooms or school laboratories.

Curriculum designers (rarely teachers, and almost never students) are typically people who have already learned how to construct these conventional similarities between

textbook language and the working language of a discipline, between classroom demonstrations and actual phenomena or working professional procedures, between teachers of a subject and those who practice it outside of schools. They can do this because, hopefully, they have actual firsthand working knowledge of professional practices outside the context of education (in the contexts of production and use). But many teachers and most students do not have such experience. Curricula which assume that they do not need it, that it is sufficient to form abstract concepts based on inherently similar textbook and classroom examples, are based on fundamental epistemological fallacies. These similarities, too, must be constructed by learned cultural procedures; they are not inherent in the instances. They are not "the same as", and they are "like" the real thing for us only AFTER we learn how to compare them TO the real thing.

Yes, humans do pattern recognition. But but we must learn to think of it not as "pattern recognition", but as pattern construction. Since we construct patterns in the context of a culture, a community, and its pervasive habits of making some kinds of meanings rather than others, we do learn to make some patterns, construct some kinds of similarities rather than others, in ways peculiar to our own community. We do this in very small steps, learning to add each new category of examples to all the others in a special new way, building up toward being able to see the similarity in all that our traditions have taught us to painstakingly construct. We do not do it in great, impossible abstract leaps. There are no guidewires of self-evident similarities to lead us only to safe landing-sites when we leap for such conclusions.

Can we learn to leap further? at least in retracing the paths made step-by-step by our cultural predecessors? (Or better think of this as a developmental recapitulation of our historical phylogeny, not driven by from the inside alone, but also by the environment with which what is inside us evolved to cooperate). Some very few of us do seem to take longer leaps and land more or less where our cultural traditions say we should. One way of understanding this is as a process of "meta-construction," in which we are guided by a learned sensitivity to the cultural habits of meaning-making that are all around us. (At least they are all around us if we are middle-class, male, etc.)

By meta-construction I do not mean what is misnamed "meta-cognition" and is really no more than self-monitoring, glorified, like self-regulation and self-discipline, by the value system of a particular, influential subculture in our society. I mean, rather, the sort of process originally envisioned (for dolphins, in fact) by Gregory Bateson (1972): that we construct patterns of patterns. Having seen how our culture constructs some kinds of similarities rather than others, according to some kinds of principles rather than others, some of us come to embody in our neurological and behavioral dispositions a successful model of these patterns, and through it we invent new similarities, on old grounds, that can be understood and accepted by others and integrated into the meaning-making practices of our community. That is, everybody thinks we must be very smart.

So, for example, when we seem to "catch on" to an abstract principle, when we guess right about how to apply a generalization to a new class of instances, we may not in fact have acquired a flexible intellectual tool at all. We may simply be following a kind of learned habit of a higher order, making what is truly just an "educated guess". When schools and teachers praise and reward those who have this knack, we are discriminating in favor of the already socially privileged, because the only way you can have this knack is by catching on to the meaning-making habits of the dominant groups

in our society and its history. If we call it intelligence, and assume it is an immutable characteristic of the individual, we will never take the trouble to teach these meaning-making procedures, step-by-step, to all the others.

At some level, most members of a community do learn, though not consciously, to construct such higher-order patterns. We can tell whether something seems culturally "alien" or not (perhaps this is even the basis of the much debated "grammaticality judgments" of native speakers about language). But the pattern-of-patterns we learn to construct are those of "our culture" in the narrow sense: the subculture we live in every day. That is not the same culture for all of us in our diverse and heterogeneous society. Even so, very few of us seem to learn to use this facility at all consciously to make new patterns, or to do what we value so much (and see so little) in education: anticipate the conventional way to make a particular pattern that has evolved in the history of our culture, before you are shown how. This facility exists, but we have misused it as evidence that what we have called abstract conceptual learning is the norm. It is almost as rare as genuine intellectual creativity, and for exactly the same reasons.

Too many of us pride ourselves that we have this facility because we have, finally, eventually, and after much struggle, re-constructed after-the-fact ways of making all these similarities seem natural. Often we achieved this only years after it was taught to us by methods which assumed that we should have easily been able to catch on, to "see" the inherent similarities that were there in front of us, so evident to our teachers, who had themselves similarly struggled for years to see them. Most of us can remember blaming ourselves for not being able to "see" abstract relations that our teachers assured us were "there" and which they were confident we "should" be able to catch on to after two or three examples. Many of us perpetuate this fallacy and its painful frustrations with our own students.

Postmodernism, constructivism, and social semiotics are not here to make our lives tough. Modernism, positivism, and abstract conceptual learning theory have already done that. We deserve a break. And so do our students.

REFERENCES

Bakhtin, Mikhail. [= V.N. Voloshinov] 1929. Marxism and the Philosophy of Language. [Cambridge, MA: Harvard University Press; 1986 edition]

-- . 1981. The Dialogic Imagination. Austin TX: University of Texas Press.

-- . 1986. Speech Genres and Other Late Essays. Austin: University of Texas Press.

Bateson, Gregory. 1972. Steps to an Ecology of Mind. New York: Ballantine Books.

Bazerman, Charles. 1988. Shaping Written Knowledge. Madison, Wis.: University of Wisconsin Press.

Bourdieu, Pierre. 1990. The Logic of Practice. Stanford, CA: Stanford University Press.

Chomsky, Noam. 1957. Syntactic Structures. The Hague: Mouton.

-- . 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press.

Cole, M., Gay, J., Glick, J. and Sharp, D. 1971. The Cultural Context of Learning and Thinking. New York: Basic Books.

Cole, M. and Scribner, S. 1974. Culture and Thought. New York: Wiley.

Derrida, Jacques. 1976. Of Grammatology. Baltimore: Johns Hopkins University Press. [Original French edition, 1967].

Foucault, Michel. 1969. The Archeology of Knowledge. New York: Random House (1972 edition).

Geertz, Clifford. 1983. Local Knowledge. New York: Basic Books.

G”del, Kurt. 1931/1962. On Formally Undecidable Propositions. New York: Basic Books.

Halliday, M.A.K. 1988. On the language of physical science. In M. Ghadessy, Ed. Registers of Written English. London: Pinter Publishers. (pp. 162 - 177).

Halliday, M.A.K. and Martin, J.D. (1993). Writing Science. Melbourne: Falmer Press.

Harding, Sandra. 1986. The Science Question in Feminism. Ithaca: Cornell University Press.

Haraway, D. 1991. Simians, Cyborgs, and Women. New York: Routledge.

Hasan, R. 1986. The ontogenesis of ideology: an interpretation of mother child talk. In T. Threadgold, Ed., Language-Semiotics-Ideology. Sydney: Sydney Association for Studies in Society and Culture.

-- . 1990. A sociolinguistic interpretation of everyday talk between mothers and children. In Halliday, M. Gibbons, J. and Nicholas, H., Eds., Learning, Keeping, and Using Language. Vol.1, pp.67-99. Amsterdam: John Benjamins.

-- . 1992. Rationality in everyday talk: from process to system. In Jan Svartvik, Ed., Directions in Corpus Linguistics. Berlin: Mouton/deGruyter.

Jakobson, Roman. 1956. Fundamentals of Language. The Hague: Mouton.

-- . 1971. Selected Writings, 2 vols. The Hague: Mouton.

Latour, Bruno. 1987. Science in Action. Cambridge, MA: Harvard University Press.

Latour, Bruno and Woolgar, Steve. 1979. Laboratory Life. Princeton, NJ: Princeton University Press.

Lave, Jean. 1988. Cognition in Practice. Cambridge: Cambridge University Press.

Lave, J. and Rogoff, B. (Eds.). 1984. Everyday Cognition : Its Development in Social Context. Cambridge, Mass. : Harvard University Press.

Lemke, J.L. 1985. Ideology, intertextuality, and the notion of register. In J.D. Benson and W.S. Greaves, Eds., Systemic Perspectives on Discourse (pp.275-294). Norwood, NJ: Ablex Publishing.

-- . 1988a. Discourses in conflict: Heteroglossia and text semantics. In J. D. Benson and W. S. Greaves, Eds. Systemic Functional Approaches to Discourse (pp. 29-50). Norwood, NJ: Ablex Publishing.

-- . 1988b. Towards a Social Semiotics of the Material Subject. T. Threadgold, Ed. Working Papers, Volume 2: Sociosemiotics (pp. 1-17). Sydney: Sydney Association for Studies in Society and Culture. See also Chapter 5 of J.L. Lemke, Textual Politics: Discourse and Social Dynamics, London: Taylor & Francis / Falmer Press (in press).

-- . 1989. Social semiotics: A new model for literacy education. In D. Bloome, Ed. Classrooms and Literacy (pp. 289-309). Norwood, NJ: Ablex Publishing.

-- . 1990a. Talking Science: Language, Learning, and Values. Norwood, NJ: Ablex Publishing.

-- . 1990b. Technical Discourse and Technocratic Ideology. In M. A. K. Halliday, John Gibbons, and Howard Nicholas, Eds. Learning, Keeping, and Using Language: Selected Papers from the 8th AILA World Congress of Applied Linguistics, Sydney 1987. Vol. II. pp. 435-460. Amsterdam: John Benjamins.

-- . (In press). Discourse, Dynamics, and Social Change. In Language as Cultural Dynamic [Focus issue of Cultural Dynamics, M.A.K. Halliday, Issue Editor]. See also Chapter 6 of J.L. Lemke, Textual Politics: Discourse and Social Dynamics, London: Taylor & Francis / Falmer Press (in press).

Levi-Strauss, C. 1963. Structural Anthroplogy. New York: Basic Books.

McCloskey, Donald N. 1985. The Rhetoric of Economics. Madison, WI: University of Wisconsin Press.

Nicholson, Linda J. 1990. Feminism/Postmodernism. New York: Routledge.

Nietzsche, Friedrich Wilhelm. 1967. On the Genealogy of Morals. New York: Vintage Books.

-- . 1989. Friedrich Nietzsche on Rhetoric and Language. New York : Oxford University Press.

Piaget, Jean. 1971. Biology and Knowledge. Chicago: University of Chicago Press. [Original French edition, 1967]

-- . 1970. Structuralism. New York: Basic Books. [Original French edition, 1968]

Thibault, Paul. 1986. Text, Discourse, and Context: A Social Semiotic Perspective. Toronto, Victoria University: Toronto Semiotic Circle Monographs.

Vico, Giambattista. 1774/1968. The New Science of Giambattista Vico. Ithaca, N.Y., Cornell University Press.

Wittgenstein, Ludwig. 1949/1958. Philosophical Investigations. Oxford: Blackwell.

TYPOLOGICAL AND TOPOLOGICAL MEANING IN DIAGNOSTIC DISCOURSE

J L Lemke

Multidimensional Meaning and Video-realism

Viewing a six-minute episode from a 'Problem-based Learning' (PBL) session recorded on videotape (see Introduction and Appendices to this special issue), I seem to see and hear several medical students and their 'coach' talking, gesturing, looking at and pointing to a complex chart, reading aloud from texts and notes, and generally making meaning by acting in and interacting with one another and the materials around them. I want to briefly analyze a few aspects of their talk and action in order to focus on how and why verbal and visual resources are integrated in this activity and how and why various kinds of evaluative meanings are getting made. The analysis is motivated in part by a concern for the materiality of the meaning-making process itself and its implications for the kinds of meanings we make and how we make them. I begin with some reflections on the surprising realism of the video medium, in which the data analyzed here have been recorded, propose an extended semiotic model for multimedia analysis, and present some highlights from my analysis of the PBL episode.

Videotape and sound-film recording of human social activity are, by the conventions of modernist culture at least, the representational technologies which most closely approximate our 'unmediated' social experience. Despite our analytical awareness that camera-angles, lighting, and editing techniques must necessarily create artifacts, and that the experience of co-presence or interactive participation (even as a silent observer) is fundamentally different from viewing a video, we nevertheless tend to construct a strong sense of 'reality' for the world on screen. A transcript of the verbal interactions in a scene is quite lifeless by comparison, even when accompanied by still images that 'storyboard' the action. Why is it that it is the dynamical integration of speech and image, on a realistic time-scale, which affects us so strongly? It is not enough to beg the question by constructing similarities to our pervasive life-experience, for that still does not tell us what there is about the time-unfolding unities of sight and sound that makes such experiences so much more meaningful to us than texts or still and silent images.

I believe that part of the answer to the mystery of video-realism is that we make meaning in social activity in two fundamentally distinct but complementary ways, which I will call the typological and topological modes (see below). Moreover, there is clearly something about the dynamical coordination in time (on the millisecond scale)

of acoustic and visual input which inclines us to integrate them functionally; that is, we make unified meanings from them, treating them as in effect unitary rather than distinct phenomena. This integration is itself achieved through the continuously variable qualities of acoustic-vocal and visual media, and so represents an aspect of their 'topological' meaningfulness.

Typological meaning is the familiar kind described by most theories of linguistics and semiotics: a material form of some sort is assigned to a culturally meaningful category, and the meaning of the category arises from its systematic contrasts with other related categories. In this view of meaning, signs, like words, are assumed to be discrete symbols; as tokens of some type-category they either do or do not possess the various criterial or distinguishing features which define one type as opposed to another. In the neat case of grammatical meaning, a pronoun is either singular or plural, masculine or feminine or neuter, first person or second or third; a verb is assigned to just one of a set of possible tenses, and so on. There are no intermediate cases between present and past, declarative and interrogative, singular and plural -- or if there are (as in some languages) they are again represented as additional discrete categories; there is no continuous variation that is meaningful. The continuous variation in the material world is reduced to categorial difference by interpreting a form as an instance of a sign. Or at least this is how signs work typologically.

But the material forms through which we make typological meaning, whether printed word, spoken utterance, or drawn symbol do always also vary from instance to instance in ways that may not be criterial for membership in a sign category, but which exhibit continuous variation that is perceivable and to which our cultures do assign meaning. I will call meanings made on the basis of continuous or quasi-continuous variation in some property of a material form topological meanings. Really, of course, these are simply two complementary strategies for making meaning, two fundamental modalities of semiosis. The unique acoustic signature of my voice saying a word, compared to other people saying that same word, identifies me to others in a meaningful way based on linguistically non-criterial differences in pronunciation that in principle vary continuously. The vocal force or emphasis at a point in my speech, and from one phrase to another, my ability to shift the timbre and sound-quality of my voice or prolong particular syllables to create meaningful nuances utilizes the continuous variability of pitch, length, and acoustic harmonics in ways that are counted as 'paralinguistic' precisely to the extent that they are not typological. The linguistic system, of course, does allow for some degree of gradations, a case of typology approximating topology, as in evaluative meanings which are inherently gradable semantically across some range from, say, 'totally impossible' through 'just barely possible' and 'really quite likely' to 'absolutely certain'. For practical purposes it is almost always possible to interpolate an intermediate degree of certainty/uncertainty between any two other degrees expressed in words, but the process quickly becomes awkward.

Language of course had to evolve to help us make sense of continuous variation in the world in such fundamental matters as spatial and temporal relationship and quantity of matter. As we developed technologies, and economies, where more and more careful quantitative distinctions were necessary, natural languages became extended semantically by the concepts and symbol systems of mathematics in order to more efficiently represent topological meaning. But from the earliest recorded times, visual representations were also in use to supplement the limited ability of language to

represent spatial relationships such as the angles of a triangle or quantitative ones such as non-simple or even irrational ratios. Visual media such as geometric diagrams or pictures sit still for us, like writing, to allow re-inspection and re-tracing of arguments, and their material extension in space allows iconic representation of continuous variation. Body gestures and communicative movement, visual signs for the observer, also help in the representation of topological meaning.

These two complementary meaning-making strategies always occur together. In speech, word choice and grammatical meaning are essentially typological, intonations and prosodies are both typological and topological, and many voice-quality and quantitative emphasis effects are purely topological. In visual representation conventional symbols and shapes often make meaning typologically, while compositional effects of relative size and placement, or meaningful degrees of variation in hue or saturation of colors are mainly topological. A gesture may function as a purely typological sign, or it may be articulated dynamically in time to also convey a topological meaning (e.g. degree of impatience). The manual sign languages of the deaf are very like speech in combining typologically discrete signs (typically hand-shapes) with topologically meaningful modes of articulation of these signs (placement, rates of movement, paths of movement) in the space between the signers.

Each of these basic modes of meaning-making foregrounds particular kinds of potentially meaningful relationships and construes 'entitities' or units of analysis as participants in these relations. Most of our conventional representational media (e.g. written text, photographs) rely far more on one of these two basic semiotic modes than on the other, or juxtapose the two after having already analytically separated them. They can at most offer us their discrete combinatorial possibilities. Video and film media, however, allow us to re-construe endlessly different inter-calibrations of these two kinds of meanings, taking multiple perspectives on how they mutually contextualize one another. While combinatorial presentations, e.g. texts with images, are discrete and static, even if presented sequentially in time, video and film combine audio and visual information, verbal speech and visible action dynamically and continuously, so that they can be experienced and made sense of simultaneously or sequentially on an unlimited number of time-scales.

I am not attempting to argue in this way for any superiority of video or film recordings as scientific data; all records have their special uses. I am trying to foreground the special importance and complexity in video-analysis of perspective-taking: critically and reflexively making explicit not only the themes and aims of our analyses, but the role played by each of the two fundamental semiotic modes, how we are inter-calibrating them, and what stance we are taking toward the multiple overlapping time-scales dynamically available to our analysis.

In the comments I wish to offer on a 6-minute video segment of the interactions among several medical students and their 'coach', I will try to focus on the relations between these two fundamental modes of meaning construction, the typological and the topological, and on some dynamical phenomena of social activity as a self-organizing process in time.

Typological and Topological Meaning-making

The participants in this video episode are engaged in a social activity in which they are constructing an approximation of medical diagnostic discourse. One student in particular (Betty) presents a number of hypotheses about the possible causes of a patient's symptoms, and other students react to these. Their tutor/coach intervenes minimally, but attempts to steer the discussion in particular ways.

I want to analyze a tension I perceive in the episode between the norms and strategies of medical diagnostic discourse as practiced by these students, and encouraged by the Coach, and the nature of the social-biological phenomena being construed with language and other semiotic resources.

The diagnostic approach, and the underlying medical terminology for events, conditions, and anatomical objects, is fundamentally, in semantic terms, a typological one. That is, it contrasts one diagnostic category with another in 'either-or' terms; it imposes a discrete terminology on continuously varying phenomena and divides even the continuous topography of the brain into bordered territories as seemingly definite as those of nation-states.

Natural phenomena, however, and both natural languages and their technical extensions, also require us to be able to take a more topological approach to making meaning with them. We need to be able to speak of quantitative and continuous variation, of multiple simultaneous and mutually non-exclusive descriptive features, of overlaps and in-betweens, of matters of degree and instability.

Our dominant intellectual culture, however, privileges the position of classical logic with its narrow view of propositions as eternally either true or false, which in turn requires typological semantic approaches to both reasoning and formal terminological systems. Sharp boundaries are required between this and that, and between true and false, contrary to the bulk of human historical experience, which shows that this is an excessively limiting way to view the world.

Natural language has evolved to provide us with resources for talking about quantity, and about degrees of certainty and uncertainty. English and most Indo-European languages, at least, do not basically treat propositions or proposals as either true or false, either good or bad. There are about a half-dozen or so semantic properties of propositions, of which Warrantability or relative probability/certainty is just one among equals (and less frequent and elaborated than, for example, Desirability). All of these are very subtly gradable in matters of degree, with a Polarity option (binary dichotomy) available in some cases through the grammar (is vs. isn't) or lexical antonyms like true, false; good, bad (though semantically these do not need to be mutually exclusive, except by cultural preference; for this analysis see Lemke 1989, 1992, 1996).

Natural languages have also been extended in those fields (linguistic registers) that have to deal critically with continuous variation and complex quantification. These extensions go, by and large, under the name of 'mathematics' insofar as mathematics is simply an extension of the semantics of natural language. The integration of mathematical and verbal reasoning is possible because of this historical relationship.

Far older still than mathematics, and also intimately involved in its history, is visual semiotics. We humans make meaning with depictional semiotic resources, ranging from our various conventionalized pictorial resources to more abstract diagrammatic and graphical ones derived from them historically. Writing systems and mathematical symbolisms represent a special case of the general unification of visual and verbal means for making meaningful representations (cf. Harris 1995 and Lemke, in press-a; Cajori 1928). Quantitative reasoning in the sciences represents perhaps the most elaborate case of integrated visual, verbal, and mathematical resources being deployed in meaning-making (see for example, Lemke, in press-b).

Natural language also co-evolved with human gestural and postural systems for communication, and indeed as an integral part of human social activity in all its material, ecological aspects. Ontogenetically, phylogenetically, and historically, speech and gesture share common origins (as do gesture and depiction; see discussion in Lemke 1994). Gesture allows us greater latitude and subtlety in making topological meaning relations than do the mainly typological resources of verbal semantics. Spatialization in gesture is akin to spatial representation in depiction.

In the segment transcribed here (see pp. xxx-xxx of this issue) we see two prime instances of the tension between typological-categorial norms in medical diagnostic discourse and the topological-quantitative nature of the biological phenomena being discusssed and constructed. One is the imposition of typologically discrete terminology on the quasi-continuous tissue manifold of the human cortex (transcript lines 15-85). The other is the imposition of the typological disjunctions of the mutually exclusive categories of medical diagnosis onto the condition of the patient (especially lines 155-192). In both cases, some more topological natural language resources, and the topological power of spatializing gestures, are used by the students to help bridge the contradiction and resolve the tension. This is what I want to examine more closely.

The Setting and the Activity

The medical school students and their coach are interacting in a "Problem-Based Learning" (PBL) session (Koschmann, Myers, Feltovich, & Barrows, 1994). The students are seated around a rectangular table in a small seminar room. From the camera's viewpoint, there is a whiteboard with the patient's case information along the left wall, and a very large free-standing Chart with sectional views of the human brain behind the opposite end of the table. The principal speaker in the episode is Betty, who sits at the near end (head) of the table, Coach to her right. Downtable left are Norman and Jenny, at the opposite end is Lill, and downtable right beyond the Coach are Maria and May (who does not take a full turn at talk in the episode).

The students are attempting to diagnose the case of a (mainly hypothetical) patient whose presenting symptoms and complete test results are available in a 'casebook' which in principle allows fine-grained diagnostic disciminations to be made, but students are expected to access this information only to test specific hypotheses (corresponding to an economical and minimally invasive approach to real patients) -- cf. the semi-humorous remark and responses at lines 307-315.

The transcribed segment begins with Betty's first hypothesis (line 12, 'my theory ...'), which leads to a digression on the exact location of the hippocampus in the human

brain. We then get a second hypothesis from Betty (line 86, 'my other theory') and a debate of the evidence for or against it, which is focussed by the Coach as an either-or question (lines 155-158). The students find the dichotomy too constraining relative to their interpretation of the case evidence, and this leads them to deploy a number of more topological meaning-making strategies as alternatives to a simple typological diagnosis. No firm conclusion is reached.

Socially, there are at least the following agendas at work:

the construction, maintenance, and negotiation of group interpersonal relations the negotiated construction of thematic views of medical phenomena the enactment of cultural and subcultural norms and formations

These agendas are intimately interwoven and interdependent as close analysis can show. They include matters of personal dominance and authority, cultural gender roles, the discourse formations of medical theory, and institutional role relationships.

Most of the action in these episodes is talk, and therefore a linguistic-semantic analysis is most revealing for what meanings are being made through this talk and how. But there is also significant use of gesture, and in the first episode use of the visual semiotic resources of the Chart is essential. Gestures enable the creation of intermediate alternatives not available in the lexicon of available diagnoses. The Chart provides representations of three-dimensional spatial relationships within the brain that are not readily described in words. An integrated analysis of at least these three semiotic modalities must be attempted (for prior work on such integration see Lemke 1987; 1995a,b; in press-b).

Locating the Hippocampus

Let's consider very briefly the first part of the episode, in which, after Betty's suggestion that the lesion causing the symptoms may be near the hippocampus, the Coach asks the students (line 28) "Where is the hippocampus?" What is worth noting here is first of all that Betty's immediate reaction is not to begin a verbal answer, but to orient to the need for "a picture" (line 29). Verbal language by itself is pretty well powerless to answer the question because its predominantly typological resources may be very good at saying what things are, but are rather limited in establishing spatial relationships, especially in three dimensions and for spatial regions of irregular shape and not readily visible location.

Norman first points to the Chart from his seat, then gets up and walks a considerable way to be able to point less ambiguously, and finally puts his finger on or almost on the chart (thus minimizing the visual ambiguity due to parallax) and traces the spatial region corresponding to the hippocampus. The information he thus conveys with the help of the Chart could not be conveyed verbally in natural language alone.

This procedure is then basically repeated by Lill for a different sectional view of the three-dimensional cortex. Semantic typology is used during the co-construction of Lill's gestural identification by other participants (go to the crevice, that's white matter, that little loop), but these expressions only work indexically (exophoric spatial deixis) together with the visual-kinesic-spatial resources being deployed here by the group to

make meaning topologically. Without the Chart's being visible to all, these locutions would be functionally useless. At one point the Coach says: (lines 58-61), "That's it. That's the hippocampus, then you go over one more gyrus and you're in the temporal lobe." His contrastive stress on temporal presses the typological approach of medical scientific terminologies. In fact, no sharp boundaries can be drawn for a 'gyrus' or a 'lobe'. The cortex is a quasi-continuous tissue manifold. Even at the microanatomical level there would not be such boundaries, but rather different cell types intermixing and overlapping in space. I am not even sure if it is absolutely possible to say for any given cell whether it belongs to the hippocampus or not in absolute terms. Nor would it necessarily be medically useful to do so.

We see in these portions of the PBL episode examples of the close functional integration of topological and typological strategies and resources for making meaning. They are clearly complementary, but there is also a tension between them produced by the typological bias of medical diagnostic discourse. We also sense here, from the videotape, the dynamics of this functional integration in time (see Rogers Hall's account of locating the hippocampus in his contribution to this volume). An even clearer example of dynamic integration occurs in the section of the episode I discuss next; there the gestural topology helps break down typological diagnostic dichotomies, while language is mobilized to express complex degrees of warrantability for various hypoptheses.

Between True and False

We turn now to a more central concern of the episode and of medical diagnostic reasoning. It is not just spatial continua that are not well represented by typological semantic strategies, it is also conditions and events. When typological categories are imposed to represent phenomena, propositions made about these phenomena in terms of such categories become problematic. It is not usually possible in life to simply say this is true and that is false, and natural language recognizes this state of affairs semantically by offering us a number of interpolations between polar truth and polar falsity. These have been analyzed in a number of ways in linguistics, most usefully in my opinion by Halliday (e.g. 1985), whose analysis has been extended in various ways by Martin (1992) and myself (Lemke 1996). My argument here, however, could easily be recast in the terms of other semantic theories.

One interpolation between true and false is that of probability. The Warrantability of a proposition, as a semantic attribute of propositions, is a matter both of degree (how likely) and of polarity (likely to be, likely not to be). We can assert or warrant a proposition both as more or less certain and as more or less uncertain. Another interpolation is that of frequency. The Usuality of a proposition is a semantic attribute we can construct for it and which tells the speaker's view of how frequent, normal, usual, expected (or rare, abnormal, unusual, surprising) it is. A not very well understood, but fairly common, extension of the semantics of Usuality (or perhaps its intersection with the semantics of Temporality), is that of Stability or Temporariness. What is not usual may also be something newly arisen, or something changed from what it has been. It may not be usual because it is only temporarily or recently the case. Finally, it is perhaps worth mentioning that there is a third systematic option in the semantics of evaluating propositions. We may do so with explicit 'subjectivity', saying for instance: 'I am sure that ...' or 'I suspect that ...' or we may 'objectify' and say instead:

'It is certain that ...' or 'It is possible that ...' and in the extreme polar case, simply: 'It is so' or 'It is not so.'

In lines 170-176, Jenny says that the condition called RIND is "somewhere in between a completed stroke and TIA [another condition]" and she makes a complex gesture co-ordinating right hand with "completed stroke" and left hand with "TIA", creating a gestural space which stands here metaphorically for the topological space of possible meanings in between the typological categories of the diagnosis. Betty then quips "like ... unstable angina of the mind!" (lines 178-181), making a semantic connection between the issue of Stability or Temporariness and the continuum of possible conditions under discussion.

These conditions differ from each other, so far as is said here, in part by a quantitative difference in how long symptoms persist. That quantitative difference can be represented spatially in contrast with the discreteness of the typological diagnostic categories, and the instability of the symptoms or condition contrasts with the implicit stability of the notion that a patient 'has a condition'. The students laugh here over the tension between a norm of clear-cut right-or-wrong diagnosis with its scientific definitiveness, and the fuzzy nature of the phenomena they must deal with.

When Jenny co-articulates her gestures and speech (including the vocal gesture of the pitch pattern for 'in be-tween', line 175) there is again a dynamic integration in time which has special force in the video that is lost in the transcript, even if supplemented by a still photo, or photo sequence, of her gesture. What linguistics and kinesics separate for analytical purposes, the physiology and functionality of human communication fuse as a unitary perceptual phenomenon and meaning-reality. No doubt they are fused as well for Jenny in the process of producing co-ordinated voicing-and-moving, and insofar as we as viewers, or Jenny's co-participants face-to-face, are entrained by the 'interactional synchronies' (Condon & Ogsten 1967, Kendon 1973) of the dynamic situation, they are fused for us as well. A focus on topological strategies and resources for making meaning cannot be separated from a concern for the general materiality, including bodily involvement, of sense-making. This applies both to the primary participants in the video and ourselves as secondary participants interacting with the video.

Returning to the analysis proper, we find The Coach a moment later (line 186) again pressing a typological view: "So which one did he have?" and the responses begin with Maria's "he's progressing to a stroke", which emphasizes Instability through the aspectual semantics of verbs (the ways we express incipience, progression, habituality, completion, etc. of processes). Norman comments: "A little bit of both", thus implicitly challenging the either/or semantics of typological categories, and the Coach's "which one?". Betty then begins (line 193) another long discussion that turns on the Stability of the patient's condition and symptoms.

In the course of this (see lines 242-245), Norman rather forcefully frames the Instability with a contrast between "we're seeing an acute leg deficit" but "now we're seeing five-over-five strength", and he makes hand gesture movements to accentuate this instability and temporariness. (Note that the issue of temporariness and change had been introduced initially by the Coach in lines 131-132: "so why do the leg findings go away?"). Typological categories stand synoptically outside of time, in an eternal

Platonic present of abstract relations. Even continuous change in properties can be assimilated to this paradigm (one of the triumphs of modern mathematics and science), but what about intermittent, discontinuous, unreliable, unstable, and merely 'temporary' phenomena or symptoms? Part of video-realism, as of life, is the meaningfulness of such phenomena. In language, they are not readily represented in individual clauses, but can be approximated over the course of longer texts or narratives. In visual semiotics again, we require a dimension of time (in production, presentation, or interpretation) to represent them.

The PBL discussion next moves on to the other main symptom, a problem with verbal language. In lines 255-260, Norman argues that the patient's speech is "screwed up", and Betty challenges this in a polar and typological way: "Is it screwed up?" Norman asserts again with a qualification "somehow" (which is a shade less definitive than the pure polar choice), and Betty concedes only in topological terms: "a little bit" and makes a gesture with her fingers held extremely close together. It is vs. It isn't has been converted again to a matter of in-between, of degree, of manner, of how much. And this is not the end of the discussion, for in her follow-up (lines 269-277) Betty invokes a whole host of Usuality resources (occasionally, rarely, often) and a construction of Instability (one part of the mental status exam vs. "the rest" of it).

By the time she gets to her conclusion (lines 279-286), the resources of Warrant-by-degree are in full sway: "I don't know" (i.e. no polar assertion, no high degree of warrantability), "I think", "would probably lean more towards" (lower degrees of probability and warrant), together with the associated Instability, "something transient that comes and goes", "at a fairly good moment".

Her final argument again turns on Instability (line 293-301), that things must have been worse at one time than they are "right now". Again her hands seem to move to show the temporal dynamics she's trying to construct, as opposed to a more static or synoptic view of a patient's definitive condition.

Conclusions

My point here is not that medical diagnostic discourse is inappropriate to the real complexity of biological and social phenomena, but that a formal emphasis on typological meaning constructions, on definitive categorizations and sharp boundaries, is necessarily in tension with the topological aspects of the phenomena. Natural language gives us some topological resources for making the kinds of meaning that are needed in such situations, and together with gestural and visual semiotic resources, as extended by mathematics and quantitative reasoning, we are reasonably well positioned to deal with them. In this episode, when the tension is strongest, the students bring these resources to bear. Both topological and typological meaning-making strategies are necessary; purely classificatory reasoning, pure classical reductio and excluded-middle reasoning is not sufficient.

The Coach here has mainly been pressing for a typological approach, and perhaps that is at times a valuable heuristic, forcing the sort of appeal to counter-evidence that also occurs in the episode. His summary comment (lines 316-320): "Some patients are vague ... don't give you the answers you wanna hear," can however be taken as marginalizing this particular case rather than emphasizing how typical it is that instances do not quite

fit general categories, putting the blame on the patient, and/or ratifying that the students should "wanna hear" more definitive answers. I am not blaming the Coach, and I may not be interpreting him generously enough. Perhaps in other episodes with this group there is more emphasis on quantitative metrics and on the impossibility of making exact mappings between continuously varying phenomena and discrete diagnostic categories. The resources the students bring to play, both semantic and gestural, seem to stand outside the official norms of the discourse, and yet they are clearly critical to making the meanings that need to be made here.

There is a great need in scientific education, especially when dealing with complex and individualized systems (e.g. local and planetary meteorology, organisms, ecosystems, etc.) to understand better the role of topological meaning in verbal semantics and reasoning, in gestural-kinesic and visual semiotics, and in the integration of mathematics with both verbal and visual reasoning. I believe that a semiotic analysis of the kind I have tried to sketch here offers some useful tools for doing this.

Regarding the video-recording medium itself, I think the points made here should be sufficient to suggest that the dynamic integration of gestural and visual communication with speech is often essential to characterizations of the kinds of meanings participants make in an interaction. It is not simply that the significance of a gesture becomes more focal when we see it co-produced with, say, speech that is constructing degrees of warrant or usuality, but that the foregrounding of such gradable meanings in the event is itself co-produced, for the participants as well as for us, by such phenomena. Meanings of many different kinds inevitably get produced in most interactions, but some kinds become more salient than others for participants, and that salience in turn sets the stage for what is more likely to emerge next in an interaction. While we may segment activity into this or that unit of action according to a particular focus of interest, and criteria based on the occurence of verbal, gestural, or other signs relevant to that interest, video-analysis still shows us that activity is continuous and that every action or sign may be construed as belonging to on-going processes on multiple time-scales, and so subserving multiple social functions. We cannot account for the dynamical, self-organizing, and emergent character of spontaneous social interaction and activity if our data, or our focus on the data, artificially dismembers the unity of meaningful action into what our various semiotic analyses (linguistic, kinesic, graphical, etc.) have evolved to describe separately. If we separate, it should only be in order to more richly reconnect.

REFERENCES

Cajori, F. (1928). A History of Mathematical Notations. Chicago: Open Court Publishing. 2 vols.

Condon, W.S. & Ogston, W.D. (1967). A segmentation of behavior. Journal of Psychiatric Research, 5, 221-235.

Halliday, M.A.K. (1985). An Introduction to Functional Grammar. London: Edward Arnold. (esp. chaps 4, 10)

Harris, R. (1995). Signs of Writing. London & New York: Routledge.

Kendon, A. (1973). The role of visible behavior in the organization of social interaction. In M. von Cranach & I. Vine (Eds.), Social communication and movement (pp.29-74). New York: Academic Press.

Koschmann, T., Myers, A., Feltovich, P., and Barrows, H. (1994). Using technology to assist in realizing effective learning and instruction: A principled approach to the use of computers in collaborative learning. Journal of the Learning Sciences, 3, 227-264.

Lemke, J.L. (1987) "Strategic Deployment of Speech and Action: A Sociosemiotic Analysis" In J.Evans & J.Deely (Eds.), Semiotics 1983: Proceedings of the Semiotic Society of America `Snowbird' Conference, (pp.67-79). New York: University Press of America.

-- . (1989). "Semantics and Social Values." WORD, 40 (1-2), 37-50.

-- . (1992). "Interpersonal Meaning in Discourse: Value Orientations." In M. Davies & L. Ravelli, Eds. Advances in Systemic Linguistics: Recent Theory and Practice, (pp.82-104). London: Pinter.

-- . (1994). "Multiplying Meaning: Literacy in a Multimedia World." Paper presented at the National Reading Conference, Charleston SC (December 1993). Arlington VA: ERIC Documents Service (ED 365 940).

-- . (1995a). "Making Towers, Making Withs" Paper presented at National Association for Research in Science Teaching, San Francisco, April 1995. Arlington VA: ERIC Documents Service (ED 384 513).

-- . (1995b). "Emergent Agendas in Collaborative Activity" Paper presented at American Educational Research Association Annual Meeting, San Francisco (April 1995). Arlington VA: ERIC Documents Service (ED 386 425).

-- . (1996). "Resources For Attitudinal Meaning: Evaluative Orientations in Newspaper Editorials." Paper presented at the American Association for Applied Linguistics, Chicago (March 1996). To be submitted to Functions of Language.

-- . (in press, a). Review of Signs of Writing (Roy Harris). Functions of Language, 4 (1).

-- . (in press, b). "Multiplying Meaning: Visual and Verbal Semiotics in Scientific Text" in J.R. Martin & R. Veel, Eds., Reading Science. London: Routledge.

Martin, J.R. (1992). English Text. Philadelphia: John Benjamins. (chaps. 6, 7)

Evaluative Meaning: The Big Picture

In social semiotics, evaluative meaning arises as an important special case of the Orientational semiotic function. The three general semiotic functions; Presentational, Orientational, and Organizational, correspond to the fact that we always make

meanings of these three kinds simultaneously. We construct or construe some presentation (aka representation) of a reality, situation, or state of affairs, real or possible. We also always orient our meaning production (talk, drawing, etc.) to some real or possible addresses and audiences, and in doing so we include some stance or orientation toward the presentational content of our meaning in the context of a system of viewpoints or attitudes and construals of reality available in our community, to us and our addressees and audiences (system of heteroglossia). And finally, we must also always construct our meanings in such a way that each meaningful whole, on each of many levels and scales, is organized by relations among its constituents parts, be they structural units or cohesive threads of semantic continuity. These social semiotic functions generalize Michael Halliday's three general meaning functions (or "meta-functions") for language (Ideational, Interpersonal, and Textual respectively).

As we make Orientational meanings, we reveal a stance or orientation toward what we are presenting: we indicate, implicitly or explicitly, to what degree we consider the presented state of affairs certain or uncertain (Warrantability), including true or false as polar extremes; we may also indicate the degree to which we regard the state of affairs as desirable or undesirable (Desirability), important or unimportant (Importance), usual and expected or surprising (Usuality), obligatory or prohibited (Normativity), mysterious or comprehensible (Comprehensibility), and serious or humorous (Humor). These seven semantic categories or Evaluative Dimensions appear to cover all the possible kinds of evaluations which can be made in English and related languages with regard to states or affairs or propositions and proposals. There are other evaluative categories for persons, things, actions, and phenomena, though the latter two overlap substantially with the seven just listed. These seven are mutually independent of one another in principle, though often conflated in practice and in the connotations of particular words in the lexicon.

Evaluative meaning plays a special role in social semiotics. It provides the link between belief and action, i.e. between semiotic representations or meanings and the material processes by which we participate in our local ecosystem. The theory asserts that we must analyze human social systems as ecological-social-semiotic systems, or networks of interdependent processes and practices: ecosocial systems for short. In such systems the flows of matter and energy and information that constitute the system materially depend critically on the meaning-making practices of human participants, individually and collectively, and particularly on (a) what we imagine to be real and possible, desirable, important, necessary, serious, etc. Our representations mediate our actions, particular in respect of our beliefs about what is and can be (Presentational meaning formations) and our values regarding what must or should be (Orientational evaluative meanings).

Evaluative Meaning Resources in Language and Visual Representation

In language there are many lexical and grammatical resources for evaluating propositions (states of affairs) and proposals (possible states of affairs), as well as people, things, actions, and phenomena. These include attitudinal qualifiers (good dog, key proposal), nouns with evaluative connotations (threat vs. promise), verbs of expression (believe vs. know), sentence and process adverbials (hopefully, significantly; possibly, frequently), and modals (could be, should be). Even more interestingly, analysis of connected text or discourse shows that evaluations in one part of the text

normally propagate, i.e. bias our interpretation of other parts of the same text. Consistency of evaluative stances also defines and links different discourse formations intertextually, so that we interpret the evaluative force of words in one text in relation to expectations deriving from the evaluative stance of related texts.

In our visual representations, we certainly also have means and conventions by which to indicate how important an element should be seen to be (Salience, or visual Importance), whether an image is to be read realistically or as a fantasy possibility (Warrantability), whether a person is good or bad, acting desirably or undesirably, ways of indicating unusualness, mysteriousness, and humorousness. In specialized genres there are very special conventions for some of these evaluative dimensions (e.g. in scientific graphs "error bars" indicate degree of Warrrantability of data points). In simple drawings, we must often rely on intertextual information, i.e. on presumptions of consistency with other visual representations, often accompanied by more explicit verbal evaluations. We will see this in the case of political cartoons.

Evaluative Meaning in Multimodal Texts

I believe that all semiosis is necessarily multimodal semiosis. That is, we always produce material signs that are susceptible of interpretation not just according to linguistic codes or meaning systems, but also according to visual ones or actional ones, etc. And as we make the signs for meanings according to those systems, we almost always also invoke interpretation in terms of verbal semantics as well. We talk about pictures as well as view them; we see and recognize conventions of typeface and font and layout as well as interpret printed words linguistically. Accordingly I believe that all the semiotic resource systems have co-evolved to work in integrated ways with one another, and that we cannot fully understand any of them until we have looked at how they make meaning together, multimodally, with one another.

I began this multimodal analysis for scientific print publications, where Presentational and Organizational meaning are much more varied and foregrounded than are Orientational meanings. When I focussed on evaluative meanings in newspaper editorials, it occurred to me to extend the analysis to include the related and saliently multimodal genre of editorial or political cartoons. These cartoons generally include verbal meaning units as well as visual ones. I have not examined these cartoons in the context of associated editorials, since for many of them such contexts were not available. My newer work will look at multimodal genres such as hypermedia on CD-ROMs and especially on the WorldWideWeb.

MATERIAL SIGN PROCESSES AND EMERGENT ECOSOCIAL

ORGANIZATION 

J L Lemke

Downward causation and the levels paradigm

Complex, self-organizing systems have traditionally been analyzed in terms of a hierarchy of organizational scales or levels. The phenomenon of "downward causation" can be interpreted as the quasi-effect of higher-level emergent patterning on the dynamics of the lower-level constituents of a self-organizing system. But the significance of such an analysis depends on the assumption that the dynamical processes which constitute the system each operate at a single characteristic scale of space and time. I will call this the assumption of scale homogeneity, and I want to question its application to social, and more generally to social-ecological systems. I will propose that when relations of meaning as well as of material interaction co-determine the dynamics of a system, we must take into account scale heterogeneity or scale-mixing as well. If system processes at very different scales are tightly coupled with one another, we may need new paradigms for system analysis and a somewhat different interpretation of the meaning and significance of "downward causation".

I want first to summarize the hierarchical or 'levels' description of complex self-organizing systems which emerged in slightly different forms over several years' discussion between Stan Salthe and myself (cf. Lemke, 1984; 1995a, esp. chap. 2, 6, and the Postscript which updates Lemke, 1984). I will then consider the role of semiosis and human-scale, human-interest viewpoints in our description of systems at all levels. To provide an example, I will focus on the process of composition of a written text, and try to explicate the multiple relevant dynamical scales and the problem of modeling them in an integrated and comprehensible account which includes various forms of 'downward causation'. Finally, I will propose some alternative models and metaphors for describing complex social-semiotic ecologies.

Many classes of self-organizing systems can be construed as hierarchical (in the sense of Salthe, 1985, 1989, 1993), where the properties of constituents provide the conditions of possibility for the emergence of higher-level organization, and higher-level patterns represent the net constraints on constituent dynamics from the interactions of all constituents in a highly-coupled system. A radical alternative to the 'levels' approach, recently proposed by Latour (1993, 1996a, 1996b), considerably opens up the discussion, and I would like to sketch out the terms for a helpful synthesis of the two positions.

To anticipate briefly, Latour proposes that traditional systems theory has too narrowly assumed that each descriptive level must maintain a homogeneous dynamical (especially temporal-spatial) scale: that all interactions constituting

the same 'level' or 'subsystem' have the same characteristic scale. He notes that many empirical investigations of human social systems, on the contrary, seem to suggest that scale-heterogeneity is the norm, and that isolating each scale on its own level makes the problem of integrating levels across scales artificially difficult. I believe that his notion of scale-heterogeneity addresses exactly the same issue as 'downward causation' but in a very different conceptual framework. In brief, I believe that his approach is more natural (but not always necessary) for systems significantly structured by the role of human semiosis and its material artifacts, whereas the 'levels' approach is more natural to elementary physical and chemical systems where scale-homogeneity is a better approximation. Organismic and ecological bio-systems fall somewhere in between, but all complex self-organizing systems which show emergent forms of order are ones for which adequate accounts must involve processes on significantly different dynamical scales. I would like to sketch our available options for performing this feat of analysis.

There is one further important point we must bear in mind in this inquiry. Emergent patterns of organization are identifiable only by privileging certain global system variables over others which a priori might equally well be used to describe the system (cf. Hasegawa, 1985). We construct variables which are relevant for our human purposes according to the cultural meaning formations in relation to which certain forms of patterning or order are salient for us. An analysis of the emergence of order in a complex system must always include the observers and their cultural criteria of meaningful patternedness as part of the system to be accounted for (cf. Bohr, 1934/61, 1958; and chapters by Finnemann, Voetmann, and Pattee in this volume). We therefore need a theory of self-organization in "ecosocial systems" (Lemke, 1994, 1995a), in which both material-interaction couplings and meaning-mediated interdependencies in the action of (human, and so also nonhuman) constituents are taken into account. The dynamics and development of ecosystems which contain humans who act according to cultural meaning criteria cannot be adequately described without a description of the meaning-systems in use and how they bias matter and energy flows in ways not predictable from regularities described solely in thermodynamic or biological terms.

These considerations will lead us specifically to the role of semiotic practices as material processes in a complex self-organizing social-ecological system, and, anticipating again, to an account of how typological-categorial distinctions, and the artifacts that mediate them, add new emergent properties to bio-thermodynamic systems which otherwise depend solely on quantitative or 'topologically' varying parameters. This account will be seen to fit very well with Latour's emphasis on the role of symbolic artifacts in the organization of human social interactions across dynamical scales. (For other examples of the complementarity of semiotic and material causation accounts, see Finnemann on artifacts and Pattee on DNA-mediated evolution in this volume.)

 

Hierarchical organization and emergent order

Let's begin with the 'hierarchical levels' account of self-organization, emergence, and downward causation. Very simply, every level of patterning or order in a complex system is seen as being characterized by units of analysis and their characteristic mutual interaction processes, each with characteristic spatial and temporal scales (within, say, the same order of magnitude) deriving from their dynamics. Taking one such level, L, as 'focal' or 'in focus' for analysis (the focus of our human concern, note for future reference), hierarchy theory proposes that two other levels are also relevant. One is the level immediately 'below' (smaller spatial, shorter dynamical time scales), L-1, where constituents of the units at level L interact with one another to produce these higher-level units. The properties of the L-1 units and their interactions provide the basis for the set of possibilities that may emerge at level L. Units at level L are self-organizing patterns of order on a larger spatial and longer time scale than what matters at level L-1. The classic examples are elementary particles interacting to form atoms, atoms to form molecules, molecules to form more complex structures ... on up through various higher levels of organization to cells, tissues, organs and organ systems, organisms, populations and ecosystems. This view presents the classic account of 'upward causation' and is the basis of the so-called 'reductionist program' in natural science: accounting for phenomena at any level by analyzing constituents and their interactions at lower levels.

Nonetheless, however neatly the 'upward causation' model fits with the general political and ideological program of modern Euro-cultural societies (i.e. individuals as more fundamental than communities), we all know that our individual behavior is still profoundly shaped, controlled, or limited in various ways by our interactions with other organisms on our own scale, and since each of these others is also similarly affected, we sum up our net mutual constraints by reference to larger-scale entities like the family, the community, the environment, society, culture, etc. In our scientific models of ecosystems we see again and again that the behavior of organisms and populations, however apparently free and independent when viewed at its own scale, nevertheless collectively repeats the same larger-scale ecological patterns and cycles.

Any entity's L-1 constituents allow it a certain range of possible behavior -- but most often we do not exercise this potential in a vacuum. We adapt to those other L-scale entities which impinge upon us, and they to still others, in ramifying chains of reaction that bind us together as communities, ecosystems, societies, cultures. As we all strive to adapt to one another, only some self-consistent collective patterns are possible for the whole swarm.

They are the patterns of emergent order and organization at level L+1, and their effects on us are just as real and material as those rising 'upwards' within us from level L-1. The constraints of mutual interaction at level L, which are summarizable as the emergent patterns of level L+1, produce, when spoken of in the pervasive linguistic metaphors of agency (cf. Pattee, this volume), 'downward causation' effects on us at level L.

While our intuitions of such matters may be sharper for our own human scale, our scientific accounts of even the most elementary dynamical systems with multiple scales (thermodynamic systems, with a micro-scale and a macro-scale) now recognize that the collective order of molar scales (entropy) emerges from correlations that accumulate over time from even the random collisional dynamics of the molecular scale (Prigogine, 1980; Prigogine & Stengers, 1984). If I am being jostled closely in a large dense crowd, I can, intrinsically, move in any direction I choose, but if others around me are tending to move in one direction, then I am more likely to be pushed from behind than from this emerging common direction, and I am more likely to encounter an open space to move into in that same relative direction -- and so is everyone else. Locally we self-organize into a micro-convection current, or align ourselves in a local 'magnetic' domain. There may be, in randomly buffeted molecules, an equal probability for reversing any particular motion in isolation, but in the crowd we are not in isolation and the net emergent dynamics of the whole is irreversible and, locally at least, negentropic. Order forms because there are only relatively few solutions to the problem of correlated motions, and when contrasted with an ideal of randomness in which all possible states of motion are equally likely, those few solutions stand out as 'orderly'.

This 'levels' model is very elegant, and when combined with the mathematical formalisms of statistical mechanics, renormalization theory, or nonlinear dynamical analysis it enables us to account for the production of order from chaos, the amplification of signals over many dynamical scales, and the emergent attractors of the dynamics of complex systems as descriptions of pattern at higher levels. In most of the cases where such models are applied in physics, chemistry, and biology the assumptions of scale-homogeneity and clearly separable levels are well justified (see Voetmann, this volume). An excellent test of this is the observation by Salthe (1985: 136 seq.) that phenomena characteristic of level L are relatively independent of the detailed dynamics of levels L-2 and below. Phenomena at level L are constituted by interactions of units at level L-1, but the very existence of these units as emergent patternings at level L-1 means that they are already collective 'averages' over the individual behavior of their own constituent units at level L-2. The L-1 units 'filter' at the same time that they organize 'noise' from level L-2, and they thereby 'buffer' level L from noise at level L-2. This is, I believe the correct interpretation of claims such as those of Varela and Maturana that

self-organizing systems are 'autonomous'. Systems at level L cannot be autonomous in the sense of insensitive to L-1 or L+1, for they are defined precisely by their selective sensitivity (filtering) to these levels. But they are relatively autonomous to noise or detailed phenomena at levels L-2 and L+2 -- unless levels L-1 and L+1 for some reason do not perform the filtering of which they are capable (see examples in Salthe, 1984: 142-3).

All this neat layering depends on scale homogeneity and relative order-of-magnitude differences in scale from level to level. Obviously effects from L-2 are already second-order negligible on the scale of level L, as well as buffered by the selective sensitivity and collective homeostasis that are constitutive of the intermediate level L-1. But 'downward causation' effects are not limited to systems for which these assumptions hold strictly.

 

Beyond levels: homogeneity vs heterogeneity of scale

Most of my human behavior, particularly what we might call my socially meaningful behavior, however much it may originate endogenously in any particular moment, and however much it may be directed to immediate interactions with present surroundings, nevertheless manages to also be quite typical and characteristic of members of my community and my period of history, and of my gender, my age-group, my social class and many subcultures to which I in some sense 'belong'. Even if we grant that dialects, social norms, epochal and community or class 'cultures' are to some extent artificial intellectual constructs produced by recording and comparing many particular momentary behaviors, nonetheless the kinds of abstract similarities and contrasts construed in these ways seem absolutely essential to any adequate account of social interaction or social organization (cf. Lemke, 1995a: chap. 2). When I speak or act in any way, not only my internal constituency and my present surroundings, but my biographical history as a member of many communities or social networks plays a part. More than this, the tools with which and through which I so often act, whether developed as part of my own body (neural nets and maps, muscular habits and skills, immune responses -- none present at birth and all specific to my individual history) or as prostheses produced by others and typical of my community (eyeglasses, hearing aids, clothing, pens and pads), or as something inextricably both (dialects, depictional styles) ensure that the history of my community, and the history of the tools themselves, are also essential for accounts of the full ecological and semiotic signficance of my moment-to-moment actions.

I write a short note to someone in my household and attach it with a small magnet to the door of the refrigerator in our kitchen. In the analysis of the

provenience, consequences, and meanings of such a culturally typical activity it is not so easy to cleanly separate different spatial-temporal dynamical scales. There is the basic human reference scale for action, on the order of seconds to minutes of performance of what counts as more or less a single coherent action or activity. Such performances typically span spatial scales of the order of the size of the human body, from millimeters of movement in the writing of the letters to perhaps meters of movement in attaching the paper to the fridge. Even in this there are perhaps three or four orders of magnitude in spatial scale and perhaps three in time. But that is nothing. The words I write may refer to recent past or future events on the order of a day or more distant from the moment of writing. The meanings of those words for me and the person I write for may well have critical dependencies on events earlier by months or years in the history of our personal relationship (a pet-name, for example). And of course the most common meanings of the words were learned by us probably in childhood, and their grammatical and semantic patterns and relationships will have histories of the order of centuries. Shifting to a spatial-extensional view of scale, personal history and community history insofar as it is relevant to the usage of words involves social networks of from several or a few dozen to tens or hundreds of thousands of people.

But even this is only the tip of the iceberg of spatial-temporal complexity of interpenetrating scales. What of the pencil and the paper without which I could not perform this culturally typical action? They too have a history, and that history implies several more scales of dynamical complexity in the social-ecological system in which my action is performable and meaningful. Where did I get them? Who made them? How and from what materials, obtained where and when? How did I learn how to use them? where and when, as part of what eco-social subsystem of cultural practices and material processes? And on what scales do they have their origins and take their present forms in relation to the specific functions they play in my present action? If we say that they are like my own muscular habits and neural memory, substrates for action at the organism level, then we need the characteristic scales of their dynamical processes to be small and fast compared to my human action scale. But they are not; some of the relevant scales exceed the human momentary-action scale. And so also for the magnet, and the refrigerator, and the architectural conventions that define kitchens and their functions and contents, and the cultural habits that make it likely a note placed there will be seen fairly soon by its intended reader. The scale of my partner's daily life routines is also somehow implied in my activity of putting a note on the fridge.

Why is this not also a problem for the hierarchical scaling of biological systems? Perhaps it is. Sometimes the long-term developmental history of particular cells may equal or exceed the typical organismic scale, but this is unusual. For tissue structures and organ systems perhaps it is not so unusual. But the functions of these constituents of the organism as a whole do not

depend on their meaning. Their significance for the organism is exhausted by their organic functioning. Insofar as their history matters at all, that history is embodied in structural form and operational behavior at the present moment. This is quite unlike the significance that attaches to the history of a word or a visual image, or even a tool. In the days before our present culture of radical commodification, the history of ownership of a famous sword, say, had very definite implications for when and how it would be used as a tool or as a symbol -- implications that could not be derived from properties found anywhere in the object itself, but which were emergent properties that existed for that sword only in the context of a much larger human community with a social history and a social memory.

Scale-heterogeneity based on embodied memory and interpretive history is definitely not a problem for elementary dynamical systems where interchangeability and the absence of non-dynamical properties or degrees of freedom ensure that there is no individuality to which a history can attach. All electrons are exactly the same; no particular one can be labelled or its history accumulated over time-scales large compared to the interactions in which it participates (cf. discussion in Lemke, 1994, 1995a: chap. 6). But as elementary units on any scale interact with many others so that collective properties emerge, we rapidly reach levels of complexity of compound units for which there can be individual identity, memory, and history -- if not yet semiotic significance (see below, and Lemke, 1994, 1995a).

The segregation of distinct temporal and spatial scales of characteristic dynamical processes is a small problem until we reach the threshold of meaning. Once meaning plays a role in material processes, once what humans do in the ecosystem depends on stories and histories, once meaning-categories influence the design of material artifacts which in turn amplify meaning's human-mediated effects on matter, then maximally different scales intersect. If we follow the connections of social practices through links attaching to both the people and the artifacts involved in them, just far enough to account for what happened here and now in terms of both material consequence and semiotic significance, then we will find that social-cultural and material-ecological processes of the widest range of scales are encountered without any clear passage from one scale-homogeneous level to another.

This point has been forcefully made by Bruno Latour in his critiques of macro-social systems theories and his interpretations of empirical studies of the role of artifacts and technologies in human cultural activity (see especially Latour, 1987, 1993, 1996a, 1996b; and studies in Lynch & Woolgar, 1990; Law 1991; Hutchins 1995; and Goodwin 1995). Latour has long argued for a heterogeneity of 'actants' (a term originally from Greimas' semiotics, cf. Greimas & Courtes, 1982) in the analysis of social practices, by which he means a more symmetrical treatment of the roles of material artifacts and

human actors, a redefinition of 'agency' away from a paradigm of independently acting agents and towards a model in which all action occurs within systems or networks of human actors and their tools, artifacts, and technologies. But recently he has emphasized that the application of his theory in empirical studies consistently argues for a 'flat' rather than a 'levels' model (Latour, 1996a, 1996b), and the core of his reasoning seems to be that the actant-networks in these studies typically have connections of and between social processes with very heterogeneous spatial and especially temporal scales.

In its current form Latour's argument is essentially that human cultural artifacts and technologies are a way of 'black-boxing' their own origins and histories, so that they can function as if they were units on a smaller scale of spatial extension and temporal process than they truly are. This is another way of saying what I have argued above: that larger-scale processes over more extensive communities and their longer-term histories are always directly implicated in semiotic artifacts, and that this fundamentally changes the scale relations for ecological-semiotic systems as opposed to more purely thermodynamic ones.

I want to argue here that both the 'levels' view and the 'scale heterogeneity' view are important tools for conceptualizing downward causation, i.e. for accounting for how larger-scale, longer-term dynamical processes become relevant for smaller-scale, shorter-term ones. Indeed I believe that a synthesis of 'levels' analysis and 'heterogeneous scale' analysis of complex systems is even more important than a reconciliation of the paradigms of 'upward' and 'downward' causation.

Salthe (1985, 1993) has made some interesting efforts to reformulate Aristotle's classic 'four causes' to fit the 'upward and downward' causation paradigm of his hierarchical levels model of complex biological systems, and he has commented there as well about the limitations of the reductionist paradigm. I would like to close this section by noting a different but related view of the reductionist program, from within a 'levels' perspective.

How is it possible in principle for a component at level L-1 of a system at level L (or spanning a number of levels) to model the levels above it? How can human organisms model the higher levels (ecosystems, cities, societies) whose downward causal influence we feel? How has it come about that we appear to be adapted by evolution to do just this?

That some modeling of higher levels is adaptive can be seen again by scale considerations. An organism has a limited capacity to sample its environment, limited in scale of spatial extension and limited in time-scales. In any given interaction on the normal human scale we sample the environment over a

period of seconds or minutes (before moving on to another interaction, perhaps in a different locale) and in a volume of a few cubic meters, or perhaps on somewhat larger spatial scales if we count distance vision and auditory input as interaction. But the ecosystem to which we belong, and which matters to our survival and the survival of our progeny, extends over much larger spaces and may have on-going processes on much slower time scales which require sampling over much longer time periods to assess. Our first asset in the struggle to know the opportunities and dangers of our habitat is memory: our bodies dynamically recreate past sampling interactions and (at least with, and perhaps even without semiotic mediation) can overlay these on top of present interactional circumstances so that we can 'see with hindsight' and act with foresight or planning. We cumulate our sampling over space and time, we comparatively register invariances and divergences.

Our second asset is communication. With symbolic resources comes the possibility of collective memory and pooling our samplings of the environment over much wider times and spaces. When the troupe travels together this effect is somewhat minimized in space but augmented in time as informational patterns are passed across age-cohorts and generations. When some members typically wander far afield and return, or when troupes exchange information, spatial extension also grows.

The key lesson here is that our means of modelling higher levels are fundamentally social and collective, and our models are built with group means (languages, vocal or gestural) and by collective actions (comparing and accumulating records and accounts). The role of writing in this activity (cf. Olson, 1994, and see Lemke, 1995b) and in the construction of other sorts of stable visual artifacts (e.g. navigational charts, scientific data archives, cf. Latour, 1987, 1990) has often been noted. Individuals do not construct detailed maps of a continent, nor could we do so without artifacts-as-records (cf. Latour's 'immutable mobiles').

In all this, we have adopted the habit of constructing the properties of wholes from samplings of their parts. Confined to the human scale in our specific interactions with the here-and-now, but benefitting from overlaying these with models of the there-and-then, we have had to learn to make sense of higher levels by piecing them together 'from below'. When this same adaptive strategy was turned to the analysis of levels below us (anatomical studies, mechanical and chemical theories) we found first that we were well-served by our technologies (our machines, built by assembling pieces into wholes), and then that we had to sample still lower levels, where changes happened too quickly for our eyes and where units were many. But we still thought in terms of aggregation and piecing together, we sampled and constructed always 'as if from below', our ancient phylogenetic trick, for which our symbolic systems

of communication and representation were themselves long adapted. We were, not very surprisingly, most successful as reductionists.

But in order to make the reductionist program work it was essential that we leave ourselves out of the picture. For once we see our representations of the levels below as actually models of our human-scale relationships to phenomena at those levels, then the neat homogeneity of scale that defines the separability of levels is broken. This paradox was of course at the heart of the great philosophical debates over quantum theory: if the data only existed because of the measuring apparatus, and so as part of the human-scale (and larger cultural-scale) experimental program, then events on our scale somehow were critical for events on the atomic scale, and vice versa (cf. the infamous lives and deaths of Schroedinger's cat). Niels Bohr's 'complementarity principle' was born from a rethinking of the reductionist program that admitted extreme heterogeneities of scale, mediated by human artifacts and technologies, in the measurement interaction itself (Bohr, 1934, 1958).

Bohr was the first to recognize that the new quantum theory presented us with a view of experience in which different interactional arrangements resulted in complementary perspectives that need not be logically consistent, compatible, or commensurable, and in many critical cases could not be so. The reductionist trick was predicated on the assumption that the different 'pieces' or views from different perspectives could always somehow be neatly fit together. But we now know that material processes cannot be comprehended, cannot be exhaustively described within any one single self-consistent formal discourse. They always overflow the limited possibilities of our semiotic models of them. It is only by building more and more semiotic-discursive models, each internally self-consistent, but not limited by requirements of mutual consistency with each other, that we can, by adding together such 'complementary' views, attain to the most complete possible account of material phenomena, including semiosis itself. Thus we still come back to a version of 'assemblage' but hopefully a more sophisticated one, one that takes into account our own role and perspective as observers, as well as the material means by which we observe, compare, and assemble -- the material mediation of our semiotic practices.

The inclusion of the observers and our technologies, and the viewpoints we embody, in the analysis of the dynamics of systems across many scales (from apparatus to electron) are also discussed in Voetmann (this volume) and in some ways extended by Bickhard (this volume) to the later field-theoretic models in physics, which are again inherently non-local -- that is, they do not necessarily respect scale: events remote in space and time become interdependent, even without any energy passing between them to communicate information. Kruse (this volume) provides us with fascinating

accounts of the historical and contemporary forms of material 'holism', an alternative to reductionism which assumes that the same sorts of patterns will be mirrored on all scales and that there is a material unity of the universe that transcends divisions into scale-homogeneous levels of analysis.

 

Materiality of Sign Processes and Scale Heterogeneity

So we are now building up a picture of material artifacts and the artifact-like (i.e. embodying culturally learned patterns) properties of our own bodies as accumulators of information sampled on larger-than-human, greater-than-here-and-now time scales, from those of the family and extended communities to those of personal biographical trajectories and wider social history. The importance of these artifacts lies foremost in their mediation of the role of meaning in human actions, actions which are always also material processes entrained in the larger dynamics of ecological-social systems. In ecosystems-with-humans, meaning matters to material dynamics because humans act in terms of semiotic categorizations and evaluations as well as in direct biological and thermodynamic response to their material environments.

Meanings are collective phenomena; their impact on the material ecosystem is large exactly to the extent that the same categorizations and evaluations are relevant to the behavior of many individual human agents. And meanings are shared to the extent of and because of common past history of interactions. At the level of the individual organism this is the phenomenon of social and cultural learning, but at the level of the community it is an emergent phenomenon, one of the small number of possible solutions to the self-consistency problem of myriad cross-coupled interactions among humans, mediated by signs and artifacts, and between humans and all material Nature. The common meanings of words, the common structures of language, the shared ideologies and genres of communication and social activity generally -- all of these emerge as characteristics of communities or subcommunities, and all of them change on time-scales long compared to cogent human moments, or even lifetimes, and as part of larger-scale ecosocial processes.

Linked through the role of meaning in the material dynamics of ecosocial systems, human organisms and natural or artifactual objects are the irreducible participants in the local dynamics of interaction. But the characteristic scales of the dynamical processes which engender them, determine how they participate in various interactions, and control how they change, develop, or evolve are often very different for organisms and for artifacts. If we construct our models of ecosocial systems by following the chain of linkages -- which agents and objects are essential here and now, which processes put them here, now, and in a form suitable for the interaction, what other agents and objects

were in turn involved as essential participants in those processes, ... and so on in indefinite regression -- then our models will be 'flat' with no distinct levels, they will be like complex intersecting networks of interactional processes with all the auto-catalytic and cross-catalytic feedback which engenders non-linear self-organization, and they will be heterogeneous in dynamical scale in the specific sense that the processes so linked will be consitutive of and constituted by systems with widely different characteristic spatial extensions and time-scales.

This view foregrounds the new feature of complex systems with semiotic as well as biological and thermodynamic processes: tight linkage across scales. But it will not serve our human purposes if it is our only model. If there are no intermediate levels of organization and order between the total ecosocial network and the human scale of organisms, artifacts, and activities, we will be at a loss to usefully model the many kinds of order that we do indeed find in language, culture, social structure, ideology, historical change, etc., or to account for these larger-scale patternings in the mutual interactions of myriad constituents at the human scale. I believe that we are adapted for survival as a species-in-an-ecosystem partly by our ability to piece together collective pictures of systems larger than our own scale of moment-to-moment living. The 'flat' models of scale heterogeneity, however, do provide for one key notion of intermediate organization: artifacts themselves.

Latour conceives of artifacts in part as 'black boxes', as material condensations of the histories and processes that gave rise to them and determine their functional potential, but at the same time as units of interaction at the human scale for which it is normally possible to ignore what is boxed up 'inside' them. If we generalize, as Latour himself usually does, from artifacts to technologies, conceived as practices in which these artifacts play a part (as tool, as product, as raw material), then we can regard artifacts as points of possible connection, points of possible relevance of the there-and-then to the here-and-now. From the viewpoint of the global actor-object-network, the dependency is very real. From the viewpoint of the human actor, the key question is whether the meaning of the object does or does not go beyond its form and function in the present circumstances. If we use the object without regard for its history and origins, i.e. for the larger systems and time scales its existence implies, then it remains a black box, an unanalyzed instrumentality. But we can always open this black box, or try to, and we do so typically when it fails to function in the expected manner as an instrumentality for here and now, or when we wish to challenge its suitability or value, or when we wish to change it for new purposes, and so on. Then our interactions with it and through it will begin to depend quite critically on larger-scale processes in which it, and now we, participate.

In this sense artifacts are the very material reality of 'downward causation' in social processes: not only the means through which larger scales of dynamical organization impinge on each moment, but also the means by which, in the production of artifacts, we both produce and model these larger scales. Latour (1987, 1990) notes, for example, that we produce artifacts such as data archives and maps as the tools by which we sum up over many sampling interactions with the environment at the human scale, and so build up more global models. At the same time, in using them, we become able to carry out different kinds of material human practices (mid-ocean navigation, global trade) that in turn alter the world on larger-than-human (as well as human) scales.

But if we also maintain the 'levels' perspective we will suspect that artifacts can hardly be unique in this respect, though they may represent an especially salient case. It is not at all clear, for example, whether language as an abstract structure, or even the sets of systematic relations in how people use language, are artifacts in the same sense as material tools or objects, but certainly they have all the same functional properties which Latour requires. Nor is it likely that one needs technological artifacts as such, since natural environmental objects can also function semiotically in addition to their biological and thermodynamic functions. Indeed the human body itself can function in this way (perhaps even pre-semiotically) insofar as it carries within itself its own history of previous interactions (cf. Bourdieu's, 1990 notion of habitus as embodied dispositions for action).

The key linkage seems to be that between matter and meaning. An artifact, a twist in the gut, or a tree outside my window may be just a tool-of-habit, just an enteroceptive sensation, just a source of shade from the sun ... or these things may also be more: the tool a shame to its maker, the twist in the gut an associative focus for a germinating and still unarticulated idea, the tree a totem of my clan planted by a particular ancestor. Material interactions and the entities we construe from them are inexhaustible sources of meaning: they overflow the terms, categories, and sets of properties that any semiotic system can assign to them. We may collectively or individually assign them only their minimal common functional meanings, or we may open them up as black boxes, or begin to pack them with links to new interactions -- including our semiotic tales about them -- even if they had none before.

Meaning is the link between matter and history; making the material meaningful potentially links the scale of humans, artifacts, and other same-scale ecological partners to the larger scales of their diverging histories and the dynamical processes that determine those histories.

 

Topological vs Typological Semiosis: Emergence of Ecosocial Systems

In order to further develop the perspective outlined in the last section, we need to see how our descriptions of ecosystems must change qualitatively when semiotic processes are at work in them. Our goal will be an account of how categorizations and valuations of material processes, by producing artifacts and other meaning-implicated material forms, transform the dynamics of ecosystems and lead to the emergence of new system properties.

The essential point is that our meaningful material interactivity in the world arises from two kinds of interdependence among specific interactional processes: a 'topological' interdependence, based on continuously variable phenomena, which is primary and characteristic of thermodynamic and biological systems without human culture, and a 'typological' interdependence introduced by those forms of human semiosis that operate in terms of discrete contrastive categories. The former tend to preserve separations of scales and allow more faithful and complete descriptions in terms of 'levels', while the addition of the latter tends to favor scale-heterogeneity. In particular, by filling the ecosocial world with artifacts designed at least partly according to typological semiotic principles, subsequent use of these artifacts in activities that further shape the world materially leads to an avalanche or cascade, a sort of chain reaction by which typological meaning colonizes the topological world. (For a fuller discussion of the 'typological' vs. 'topological' distinction, see Lemke, in press-c.)

We need, most basically, an understanding of semiosis as a material process in an ecological system. Our own cultural traditions in the centuries since Descartes have too radically disjoined the material and (under the older name of the 'mental') the semiotic. We have one set of discourses for talking about matter in the languages of physics, chemistry, and biology, and a completely different set for talking about meaning in the languages of semiotics, linguistics, and cultural anthropology. Yet we know that every sign has a material phenomenon as its representamen (sign-vehicle, signifier, carrier), that every process of semiosis is not just a social and cultural practice, but also a material activity in which not just humans but also non-human elements of the ecosytem participate.

From the ecological standpoint, we know that when an ecosystem contains a human society, we cannot account for the dynamics of the total system unless we take into account the beliefs and values of a human culture. Which trees are cut, which crops are cultivated, what kinds of raw materials transported where, depends not simply on the physical, chemical, and biological properties of human organisms or other components of the ecosystem (biotic and abiotic), but also on the cultural values assigned to actions, constructions, and objects. These phenomena depend on the beliefs of a community, depend

on purely cultural customs. These customs must be consistent, in some broad sense, with the other material aspects of the ecology, but there is still such broad latitude for differences of culture, for differences of meanings made that affect matter moved, that we cannot hope to account for the changes, for the total dynamics and trajectory of such an ecosocial system unless we take culture and semiosis as well as physics and biology into account.

What kinds of material systems can support semiosis? This will depend on how broadly or narrowly we define the making of meaning. If we take the broadest possible definition, the most inclusive one, then we have the opportunity to examine how semiosis itself has evolved with the processes of self-organization and complexification of the cosmos (cf. the similar project of C.S. Peirce who saw the processes of human semiosis as continuous with, and a veritable extension of the general tendency to self-organization in the evolution of the cosmos, which he referred to as matter's propensity of 'habit-taking' ; Peirce, 1992: chaps 19, 24). We can also examine the degrees of complexity of various classes of material systems with an eye to imagining how close they might come to what we would be happy to call semiosis sensu stricto.

I have tried to do this following Salthe's notion of a specification hierarchy (Salthe, 1985, 1989, 1993; Lemke 1994, 1995a: chap. 6). This is a formal scheme of nested classification in which each class of system is a subclass of the previous one, from an outermost class of systems whose dynamics can be adequately accounted for with the fewest assumed properties or characteristics, in the simplest discourses with the fewest number of primitive terms, to those which successively require more complex descriptions, adding to the descriptive apparatus needed for the less specified or more generic systems further properties which are newly relevant. This model would map smoothly onto an evolutionary model in which systems of greater complexity arise from systems of lesser complexity by successive differentiation, and by the emergence of new properties by processes of self-organization and symmetry-breaking within a matrix system possessing only the more general properties. Whether cosmological evolution, seen from the human viewpoint, follows the sequence of a specification hierarchy or not, the latter is still a very useful way to formulate the degrees of complexity in observable types of systems, and this suits our present purpose very well. (Emmeche et al., this volume, seem to assume that historical emergence follows the specification hierarchy, but I believe this view needs to be somewhat modified from an 'ascent' to ever higher levels to a progressive interpolation of ever more specified kinds of systems between less specified ones above and below; see Afterword below.)

The simplest dynamical systems we know are the ones I will call elementary dynamical systems. They are typified by the elementary particles of physics

and their interactions. For purposes of describing their possible dynamical participation in an interaction, quantum theory requires that only a small number of properties be specified. Moreover, it appears that these systems can have no other properties than these essential dynamical ones. The number of degrees of freedom, the number of ways in which such a system can be identified by its behavior, is completely exhausted by the properties necessary to account for its fundamental interactions. For this reason, modern physics says that electrons have no individuality, no history, no culture that matters to their potential behavior under any and all circumstances. Every electron in the same quantum state will behave the same way with the same probability as any other electron. When two electrons collide, one cannot trace their identities from before their interaction to after it. Each could as well have taken the role of the other, and the experimental results agree only with this otherwise strange assumption.

Electrons and other constituents of elementary dynamical systems are truly simple, truly minimal in their complexity. They are not really even individual entities in the sense we expect for macroscopic systems like ourselves. Electrons and atoms do not age. They have no history, no individuality, no youth, maturity, or old age. They are generic, and their science is a science of the generic.

How complex, and complex in what ways, must a system be to show the history, individuality, and diversity characteristic of cultures and meaning-making systems? We know from the work of Ilya Prigogine (1961, 1962, 1980; Prigogine & Stengers, 1984) on irreversible thermodynamics and complexity theory that statistical ensembles of elementary dynamical systems begin to have histories that matter. They are systems for which we need to define an entropy. They break the time-reversal symmetry of elementary dynamical systems. Macroscopically they are still not truly individual, though microscopically perhaps they are. They do not yet have individualized macro-developmental trajectories. They suffer the irreversible effects of history, they are 'in time', but they do not yet have unique individual biographies, trajectories which matter to their dynamics. This further step is taken with the next more highly specified class of systems in our hierarchy, the so-called 'dissipative systems' which export entropy to their environments and feed on the order in those environments (on the maintenance of a gradient between system and environment) in order to self-organize, to increase their internal dynamical and morphological structure, to develop along a trajectory typical of their system type, from one dynamical regime to another. Such systems as flames and tornadoes, hurricanes and Rayleigh-Benard convection cells (see Table 1).

Table 1. Specification hierarchy nesting ecosocial systems

 

System Types/Classes Additional Properties

 

Elementary Dynamical Systems(electrons, atoms, small molecules) Energy, mass, identicality

Macroscopic Systems with Irreversibility(paper clips, balloons, water droplets) Entropy, memory, aging, identity

Dissipative Structures = Dynamic Open Systems(flames, dust-devils, hurricanes)                        Emergent organization, individuality, developmental trajectory

Autocatalytic Self-Organizing Systems(Cairns-Smith clays, Eigen-Schuster hypercycles) Autocatalytic-crosscatalytic interdependencies

Epigenetic-Developmental Systems(Salthe dust-devils, ...) Recapitulation of evolvable type-specific trajectory

Genetic Evolutionary Systems(Ecosystems > organisms; A-life configurations) Recombinant, transferable genotypes

Ecosocial Systems(Ecosystems-with-cultures > semiotic practices-with-persons)

                             Meaning-construal-dependent material activities

-------------------------------------------------------------------

 

These systems develop, but they do not conserve the information acquired through interactions with an environment which shapes their development, nor transmit it to future generations. There is as yet no epigenesis, no evolution. There is, however, already one feature which will later prove crucial to our analysis of semiosis in ecosocial systems: the dynamics of the system as we ordinarily define it cannot be defined in terms of processes strictly internal to the system: the dynamics is always transvective, it always crosses the boundary of what we call 'the system'. Indeed the maintenance of the structural or dynamical integrity in terms of which it is possible for us to define it as 'a system' depends directly and critically on processes of exchange of matter, energy, and information with an environment. In this sense while it

may be defined as an individual, it is not in any sense autonomous. It is, in fact, merely an isolable component of a larger dynamical system. It is always a subsystem, and to understand its dynamics, we must always examine the supersystem of which it is an integral part. It cannot exist apart from its participation in this supersystem.

The properties of each class of systems we are describing are inherited by all the subsequent subclasses of the specification hierarchy (see Table 1). A dissipative system is a thermodynamic system, a thermodynamic system is a physical system with the same parameters and degrees of freedom of elementary dynamical systems. As we shall see directly, an organism is a dissipative system, and so is an ecosystem. Indeed they are two levels of organization in the same system, and that system belongs also to the class of dissipative systems, and to all the classes above it in the specification hierarchy, a fortiori.

Where are we headed in this sequence of classes of systems? Obviously the concentric circles of subclasses of more highly specified types of systems is converging on the point from which it is being drawn: on the cultural systems in which humans and their ecologies make meanings about classes of systems. Along the way, we hopefully will gain some further insights into what makes meaning-making possible in a material system.

An epigenetic system is a developing system that recapitulates the major stages along a developmental trajectory typical of its kind. It is a system that develops according to its kind, recapitulating a sequence of bifurcations in its dynamics that may have evolved over many generations of its predecessors. I hope it is clear that while we have for some time now been using the language of living systems, that at no point in the specification hierarchy that we have been defining (complex systems with irreversibility, dissipative structures, developing systems, epigenetic systems) is there a clear transition to Life, as such. Hurricanes are alive in many significant ways; so is the Planet as a whole. Organismic life as we know it is based on a very specific strategy (DNA-mediated epigenesis), but ecosystems are also alive and use a different strategy. What is special about the class of epigenetic systems is that the developmental trajectories of individuals recapitulate a prior evolution of the trajectory of their type.

How is recapitulation possible? Epigenesis further specifies the nature of development: epigenetic development is development guided by an environment which is approximately the same for different individual systems and which changes relatively slowly compared to the lifetime of these systems. The sequence of bifurcations, of development, cannot be left entirely to chance, to random fluctuations, if there is to be recapitulation. Random fluctuations must be harnessed and guided by an external source of

information, regulation, and control, and that can only reside in the environment of the developing system. An adequate analysis of a developing system must not only be extended in time, it must also extend beyond the system itself to examine system-environment interactions: it must extend to the immediate supersystem that contains both the system under focus and its immediate environment (cf. Lemke, 1984, 1995a: 159-166).

In epigenetic systems, a new bifurcation in the developing dynamics of an individual leads to effects on the environment that favor similar bifurcations in other individuals. A series of "accidental" dust-devils in a narrow defile might erode landscape surfaces in a way that produces contours which favor the formation of other very similar dust-devils in that same place (cf. Salthe, 1993 p. 42-43). Globules of organic polymers in a tidal pool, engaged in autocatalytic chemical reactions (i.e. proto-life), might modify the surrounding silicate clays (their external, proto-DNA) in ways which tend to favor recapitulation of their latest chemical innovations when future globules develop in the same pool. In each case, along with epigenesis comes a supersystem (dustdevils-plus-landscape, globules-plus-clays-in-tidal-pool) and a hierarchical relation of system and supersystem. That hierarchical relation is one of scale (cf. Salthe, 1985, 1989, 1993, who clearly distinguishes scale hierarchies from specification hierarchies), in which the supersystem is more stable, changes more slowly, and exerts a regulatory influence on the dynamics of the now "sub" -system. In the case of organismic lifeforms, the relatively stable "environmental" molecules (RNA, DNA) were eventually internalized, incorporated into the supersystem which became the modern cell.

But epigenesis depends only on a system's being integrated into a supersystem which can in turn regulate the subsystem's development. It depends only on the possibility that innovations by individual subsystems can be recapitulated because information about them (or leading to them) is stored in the long-term "memory" of the supersystem environment. The DNA strategy of organismic life is only one specific way in which this can happen. Epigenesis is simply development under an environmental guidance that enables recapitulation of type-trajectories in individual development. (For a more complete picture of how development and evolution are linked by DNA, including the complementary roles of 'typological' , semiotic constraints and 'topological' dynamical ones, see Pattee, this volume. For a complex account of subsystem-supersystem relations and the interplay of type-specific equifinality and individuation along particular developmental trajectories, see the account of language acquisition in Hirsh-Pasek, Hollich & Tucker, also in this volume.)

My simple account of things (like many accounts of the origin of organismic life) is a bit backwards: there have always been supersystems, there have always been ecosystems, there has always been a planetary dynamical system. Particular self-organizing units always came into being in the context of

supersystem environments. Life did not begin with micro-organisms that eventually got together to form ecosystems that eventually united into the living planetary system ("Gaia" after Lovelock, 1989). There was always Gaia, even before organic life, and there were always the chemical, atmospheric, oceanographic, and geological precursors of biological ecosystems. What has happened in the history of the planet is that new intermediate levels of organization have emerged between the total Gaia-system and her molecular subsystems (cf. the discussion of this issue in Moreno & Umerez, this volume). Ecosocial systems and the human cultures they sustain form one of those intermediate levels. These levels of organization, each on a different scale of physical size and mass, rates of change, energy transfer, etc. are (partially) regulated by their integration into the larger ones that contain them, and in turn (partially) regulate the smaller-scale ones that they contain.

All epigenetic systems belong to regulatory subsystem-supersystem hierarchies of this kind across a range of scales from the molecular to the planetary. At or near the human scale, organismic lifeforms are not the only epigenetic systems, there are also ecosystems.

Ecosystems are individuals. Their biographies partly recapitulate during ecological succession (Odum, 1983; Schneider, 1988) the trajectory of ecosystems of their specific type. Unlike organisms, ecosystems do not seem to die, but to undergo continual processes of local decay, replacement, and variable succession, resulting in a whole, a supersystem which is a mixed-age aggregate, a mosaic of ecological patches, each of which is itself an individual on a smaller space-time scale. Human communities exist as patches, and networks (see below), within natural ecosystems. Our communities are parts of larger ecosystems, and even our most artifactual cities show all the properties of ecosystems as a class. Any architectural survey will show the mixed-age mosaic, the mixed-use patches, the local diversity of 'species' types (person-types, artifact-types, natural types), in intimate dynamical interdependence.

But the dynamics of such an ecosocial system (Lemke, 1994, 1995a) depends not just on the volume and biotoxicity of wastes, the nutrient needs of the population, the structural properties of building materials, the available arable land; it depends also on cultural food preferences, on building styles, on technological histories, on political structures and social values. It depends on the activities by which humans not only move and transform matter and energy but also assign value and meaning. It depends on the critical link between activity and language (cf. Vygotsky, 1963; Leontiev, 1978), action context and meaning.

How can we describe this ecologically? Human networks of activity, like many of the dynamical subsystems of an ecosystem, are not strictly space-time localized 'patches'. As Latour has emphasized and his co-workers shown in so many cases (Latour, 1987, 1993; Law, 1991), networks of activity have a different topology from localized subsystems. I discuss this in more detail below (see also Lemke, in press -a), but the critical point for our purposes is that humans and the non-human species and material forms that co-participate in ecosocial and cultural practices and processes within a network interact over long distances and even at considerable removes in time more intensely in many cases than they do with objects and persons close at hand but not in that network. Our transport networks, our information and communication networks, our economic trade-routes of exchange, have always had a 1-dimensional reticular network topology spread through a three-dimensional ecosystem. This is true also of the food-webs and carbon-exchange cycles even of ecosystems that lack human participation.

One of the functions of our network subsystems, one of the kinds of activities that takes place in them is semiosis. But we have in mind now not semiosis per se in some idealist sense, not some mental processes with no consequences outside an immaterial 'mind'. We are concerned rather with a view of 'cognition' more like that of Gregory Bateson (1972) or of the situated cognition models of the last decade (e.g. Lave, 1988; Kirshner & Whitson, 1997; Lemke, 1997), in which the material substrate of semiotic processes extends always beyond the organism, is always in fact a process characteristic of the supersystem, and not internal to an organism (cf. Smith & Thelen, 1993; Lemke, 1996). We are not concerned with 'thought' that does no work in the ecosystem, because there can be no semiotic process uncoupled to the material systems which are its dynamical basis. Every 'thought' is part of a material activity, and its form and its consequences depend on the material systems and processes through which it occurs. These are only partly neurological (and neurohumoral) processes: they are also active, efferent and motor processes. From perception to memory to reasoning, all human neural activity includes both afferent and efferent connectivities (even if pre-emptively inhibited and/or re-entrantly diverted, cf. Edelman, 1992), and most human meaning-making occurs in the context of immediate motor activity. Our perception is the product of our action: the Umwelt made by our specific way of participating in the ecosystem (cf. Gibson, 1979 on affordances; von Uexkull, 1926; Smith & Thelen, 1993).

We need to represent semiosis as an integral part of activity in an ecosystem. Integral in that it is engendered by such activity, evolved to function as part of such activity, is shaped moment-to-moment by the activity, and has its consequences in the activity. In Bateson's famous example (1972: 458), the chain of differences that make a difference, which constitutes cognition or semiosis, is one aspect of activity itself, and its moment-to-moment trajectory

derives from the loop of action (hefting an axe), consequence (interaction of axe and tree), feedback (perception of recoil), and modification of action (the next swing). However ethically questionable this murder of trees may be, it readily situates itself within the larger activities and networks of activities of an ecosocial system. The axe is swung initially as part of a larger activity (tree-killing) which depends on an economy, on cultural values for certain kinds of wood, on a technology (for use of the wood as well as for cutting the tree), on the people who will buy the wood, those who will use it, those who forged the axe-blade, and so on. Semiosis is always in medias res, plunging into the midst of events, at once material and social-cultural. There is no 'mind' outside of an 'ecology' which makes it materially possible and culturally meaningful. (Bateson's title is Towards an Ecology of Mind in this sense.)

Meaning-making is a material process in a material ecosystem. Its forms have evolved and are dynamically shaped from moment-to-moment as aspects of human-mediated activities in network subsystems of ecosystems. Materiality is as fundamental to an understanding of semiosis as is social function. Indeed these two are inseparable in a model of ecosocial dynamics. Every social practice is also a material process. As social practice it has semiotic relations to other social practices, construed by the semiotic activities of human communities. As material process, it participates in material, eco-bio-physical interactions with other process of the ecosystem. It is this double-connectedness that gives to ecosocial systems their enormous increase in complexity over other ecosystems. There are so many more possible couplings of processes/practices through the mediation of semiotic relations as part of the activities of humans (and non-humans) in the system.

Finally, we can begin to characterize the materiality of semiosis in two complementary ways (see also Lemke, in press-c). In the first, more general perspective (analogous to Peirce's, see above), which I will call the 'topological' one, semiosis arises in any self-organizing system to the extent that there are differences that make differences, and that what difference a given difference makes in turn depends on some other feature of the system (for this 'meta-redundancy' view of semiosis, deriving from Bateson, see Lemke, 1984, 1995a: 166-174). This kind of semiosis is quantitative. It need not depend on categories or contrasts, it rules in the domain of the analogue and the continuously varying -- a domain in which all material systems are situated, including our ecosocial ones. The mathematical descriptions of classical physical science developed precisely to describe this sort of quantitative covariation.

But mathematics itself initially grew out of a very different sort of semiotic resource: language, for which meaning arises by discrete (not continuous) covariation of categories and types. Classical semiotics, deriving from

Saussure (1915/1959, but see Thibault, 1997 for a re-appraisal in relation to dynamical systems and topological semiosis), foregrounded the principle of valeur, according to which the meaning of a sign is a function of its place in a system of contrasts with other signs. Which other signs it is relevantly in contrast with depends on the wider context of its occurence, a fact which leads again to Bateson's differences that make a difference in a fully contextual and relational model of semiosis (see Lemke, 1984, 1995a: Postscript).

This is perhaps not the place to attempt an analysis of how typological semiotic practices can arise from a topological substrate; certainly it seems plausible that this phenomenon is itself akin to self-organization in complex systems, to the emergence of attractors of the dynamics of a 'topological' system which then bifurcate into two regimes, each of which stands in potential 'typological' contrast with the other. Of more relevance to our immediate concerns here is what happens when human collective interaction in the ecosystem becomes self-organized in such a way that linguistic and other semiotic categories play a role in human material actions in the larger system.

There are two particularly important cases, I think. One is the role of human semiotic valuations in determining what species we favor or disfavor, what materials we accumulate or disperse, and the differential ways we treat various categories of our fellow humans that are also grounded in such valuations. The other, coming full circle to Latour's arguments and the foregoing extensions of them, is the role of human semiotic categories and categorial-conceptual reasoning in the design and engineering and modes of use of our 'artifacts'. Whether we merely reshape the 'natural' environment (foraging, primitive gardening) or construct more completely artificial 'kinds' (mechanical and electronic devices) with no prior history in the ecosystem, we are providing the material means for still further human activity predicated on the use of these artifacts, and in most cases predicated also on the adoption of the categorial-conceptual logic of their functioning and use (if not always also of their design and production). Each new artifact, from a myth or a speech genre (cf. Bakhtin, 1953/1986) to a container or a computer, enables patterns of human activity in the ecosocial system that tend to multiply and project typological meaning into other domains. Typological semiosis is contagious, and artifacts are among its primary vectors.

One very simple way to appreciate the pervasiveness of the consequences of this chain reaction in which typological distinctions in one domain or material medium beget typological differentiations in others is to consider how insulated from the systems of cultural categories a child growing up in our ecosocial system could remain? Even in the absence of formal education or explicit family or peer tuition in such matters, merely as a result of operating the pervasive gadgets and technologies of daily life, categorial culture comes

with participation in material-artifactual culture. The typological, conceptual-categorial dimensions of our culture are built into our artifactual worlds, inescapably.

 

Textproduction: Linguistic Technology and Scale Heterogeneity

Are words artifacts? Certainly documents are, because of their obvious materiality, artificiality, and the role they play in our technologies of communication and representation beyond the scale of the immediate here-and-now. Spoken language is rather more evanescent, especially inner speech, which verges on thought itself. But materially the spoken word is a cultural modification of a natural feature of the ecosystem: of breathing and grunting, of cries of pain and danger, of vocalizations more topologically determined. Presumably for long periods prior to the spread of literacy technologies, oral traditions played an equivalent role to documents; and spoken dialects are examples of emergent features of on-going collective interaction, relevantly describable only for scales of systems well beyond the individual organism or the momentary dyadic interaction. The dialect we speak very certainly shapes or constrains the kinds of things we are likely to say or think -- one of the most powerful of all the examples of 'downward causation' (for this neo-Whorfian view see for example Hasan, 1986, 1990, 1992a, 1992b; Silverstein, 1979, Lucy, 1992) .

Language is a phenomenon of seemingly inexhaustible complexity, and every linguistics has had to pay attention to only some of its manifold aspects. For present purposes it is perhaps best to think of language as the sum total of all resources for verbal meaning. That includes not just the words themselves and the typical grammatical patterns that link them into phrases, clauses, and sentences, but their systematic semantic relationships (synonymy, contrast, hyponymy, etc.), the typical ways-of-speaking about various topics in a community (thematic formations, speech genres, discourse voices -- see discussions in Lemke, 1995a, in press-a, and references therein), and even the typical stories told in those ways (cf. Threadgold & Kress, 1988; Lemke, in press-b). It also includes the meaning shifts associated with the intonations and pacings with which we speak the words ... and much else.

Now here is our central mystery: How does it happen that a particular writer (or speaker) produces, word by word, and sentence by sentence, a text, that is not predetermined in detail by any explicit plan (perhaps only by some general goals or an on-going activity of which the writing forms a part) -- and which indeed quite often actually surprises us when we see what we've wound up having written -- and yet, in almost every case the resulting text can be seen as quite typical of a particular genre of a particular culture and subculture in a

particular historical period? For all the creativity and indeterminacy of the process of textproduction itself, certain 'constraints' of genre and discourse conventions will nonetheless supervene from larger scales of the ecosocial system. (For a discussion of the historical evolution of written genres through their dynamic integration into larger social-political systems, see Andersen, this volume.)

If we open the black box that is the finished and completed document, we expose the processes of its production: the other times and places, the other participants in the larger-scale systems of text production as a process. This reminds us that documents are not produced solely at the human here-and-now scale we imagine, that the relevant system in which production as a total process takes place is not limited to this scale. Somehow the larger-scale social systems are speaking through us, as in Bakhtin's metaphor of 'ventriloquation'. Bakhtin (1935/1981, 1953/1986) offers us a germinal insight into these processes in his more general notion of the pervasive 'dialogicality' of language in use. Every word, every expression form, is something that we appropriate from another, and with it comes a history, a collective memory of its uses in others' mouths and texts, which fills out its connotative meaning and cannot be ignored in any use we make of it. No word is entirely our own, and the richness of meanings in the words we appropriate depends precisely on their partial otherness.

From Bakhtin's basic insight Kristeva (1980) and others (see references in Lemke, 1985, 1995a) developed the notion of 'intertextuality'. As a principle, intertextuality reminds us that the meaning of each use of language here and now depends in part on how we connect it to other uses of language there-and-then. The relevant 'intertexts' of any given text are those which echo in its meanings as we read and interpret it, but they are also those whose resonances contributed to the original selection of words and form in our production of the text. (Naturally many texts only imply and do not explicitly state, in footnotes or by citations, the intertexts of their production, so that interpretation may bring to bear still other intertexts, resulting in further meaning possibilities.)

As I write I consult my notes, another text. I switch windows on my screen to view other papers I have written, and bibliographies, and half-formed essays to be incorporated here. I have stacks of books and papers by my side to which I refer, and some I cite for you here and some I do not. I recreate in memory the outlines and key expressions of still other texts I have read, and I do so under the stimulus of some of the materially present intertexts I have before me, as well as of my re-readings of this very text as I have written it so far. Later this first draft of my text will become an intertext for future drafts.

In all these ways the extent, the spatial and temporal scale, of this present text-production exceeds the immediate processes of here and now. But there are also more intermediaries of an artifactual sort. In addition to fully formed texts here present and remembered highlights of texts past, there are also the accumulated dispositions of my body and its memory, my 'know-how' for writing this academic style, for shaping the conventions of this genre, and less consciously, for writing my dialect of English in this historical epoch. Do I write as a man? in a masculine manner? with the semantic orientation of the upper-middle class? with the interests and emphases of an academic trained as a physicist? widely read in linguistics, ethnography, semiotics, philosophy? And can we say simply that all these dispositions are solely to be regarded as operating on the scale of my organism in isolation, as Bourdieu (1990, 1991) sums them in his notion of habitus? Yes, my bodily processes play critical roles -- but insofar as I-as-writer am concerned, the agency of this writing is distributed more widely across a larger material system and longer time-scales than any which are characteristic of me-as-organism.

I cannot write without keyboard, display screen, and all that mediates between them, or without motor-facility with pen, and paper of the right sort. My writing is as much in my fingers as in my brain, for the feedback loops between them and my eyes' sight make it no longer possible to ascribe exclusive originary agency solely to one part of this integrated system (cf. Bateson's tree-and-axe loops above). Here the differences that make a difference, however, are not simply perceptual forms and tactile feedback as simple stimuli, but rather, American radical behaviorism notwithstanding, in human action and all the systems in which humans participate, their meaning matters as well. I do not write from a preformed plan straight through to completed text. At each step I read what I am writing, read back what I have written, and new associations are made. Meanings reinterpret what the words say, differently or more richly perhaps than they meant as I first wrote them; these meanings evoke still other intertexts, still newer meaning possibilities for re-writing and editing what I just wrote, and for writing a new next phrase or sentence that I would not have planned to write before (see Lemke, 1991). My own text as growing artifact before me also has co-agency with my body and brain in this activity, not simply material co-agency, as the black boxes hidden in the writing loop do, but semiotic co-agency as well: contributors to developing meaning. The system in which meaning is being made is the larger supersystem. (Togeby, this volume, seems to suggest that text is analyzable in terms of fixed sets of goals; this may indeed be possible and useful, but I do not believe text is in general produced in this way, even though many of us probably believe it should be.)

My partners in writing include all those material intertexts I have already mentioned, and through them their agents of production, all the remembered and imagined texts, and the very dispositions to write in particular ways and

utilize particular themes and ideologies, that enter in through my body's writerly habitus. I am saying very literally that my internal organism cannot be the sole author of the texts that get written by my hand, that textproduction is a process across all these system scales. The text is not being produced solely as part of a process on the scale of a meter or so around about me and over times of the order of each attack on the keyboard or the composition of each sentence or paragraph. The text is also being produced as part of processes on much larger scales that operate over much longer characteristic times -- as part of the evolution of a discourse formation or a genre in a subcommunity (cf. Andersen, this volume), as part of the evolution of a dialect. Textproduction belongs also to community- and ecosocial system-scale processes, as well as to organism- and artifact- scale processes. (See also in this connection Togeby's insightful discussion in this volume of textual coherence as a phenomenon across time.)

This rather radical picture of a process like textproduction has an even more radical implication for human consciousness. Textproduction is a special case of meaning making, and so of what we loosely call 'thought'. Together with the more topological material ground of consciousness (our being-in-the-world interactively, much of which is also scale-heterogeneous), this analysis implies that meaning-making consciousness need not be considered as a process confined solely to the organismic scale. Our meaning-awareness, and perhaps a good bit of our primary awareness prior-to-meaning, are also aspects of larger-scale on-going processes. Consciousness is a cross-temporal phenomenon; it exists on multiple time-scales simultaneously, and some of those scales may be very long indeed.

 

New paradigms for the study of complex systems

What I have proposed here is an effort to extend our usual paradigms and metaphors for understanding complex self-organizing systems from those which assume strict scale-homogeneity of levels of organization to ones which also allow us to think, when needed, in terms of scale-heterogeneity.

A great deal of further discussion and elaboration of these issues will be needed. I believe that it will be very fruitful to examine, across the widest range of different kinds of empirically researched systems, the extent to which the assumption of scale-homogeneity is justified and the extent to which it is helpful to supplement it with views that emphasize multi-scale processes and strong cross-scale linkages. I would like to end here by mentioning a useful, if slightly simplified, set of metaphors for thinking about levels and scale-heterogeneity.

Our classic view of hierarchical levels in general systems theory models each level or dynamical domain on a sphere: a three-dimensional region with a definite spatial-extension scale, in which processes on this characteristic scale take place with some corresponding characteristic time. Our view of subsystems at lower levels models them as spheres or three-dimensional regions of smaller spatial-extension scale within the larger spheres, and so on up and down the hierarchy. There are two intimately related assumptions in such a prototypical view: (a) that there is a single definable spatial and temporal scale for the processes identifiable within the sphere, and (b) that points which are nearer in space, and events which are nearer in time, are more likely to be linked by interactions, more likely to be co-participants in larger events, more likely to be constitutive components of the next larger system level, than those which are remote (and so far outside the characteristic sphere).

But this view fails to some extent even in non-artifactual systems. Two ecological zones within the same watershed, or riparian zones along the same river, or regions within the influence of the same oceanic current, are more likely to interact with one another than they are with many points that are nearer in three-dimensional space but are not linked by these natural quasi- one-dimensional networks. Pollution dumped into a stream may have effects far down-river but none a few meters inland from the dumpsite. The relevant scale here is the long distance along the one-dimensional network link, and not any three-dimensional sphere with that (or any other) characteristic spatial-extension scale. Artifactual examples include railway and road transport networks, and telephonic and signal-cable system nets (including today the global Internet). I am far more intensively in interaction (for some purposes) with individuals in Italy, Denmark, and Australia than I am with most neighbors in my own town or street. From the viewpoint of a scale-and-levels model of the 'spherical' type, these network artifacts produce scale-inhomogeneities in processes which are both local and global.

More generally there are always a number of different co-dimensional manifolds that can be embedded in a space of N dimensions. In 3-dimensional space, in addition to three-dimensional regions, typified by the interior of spheres with a single spatial scale parameter, there are also what we will call 'networks', which are of co-dimension one. They are reticula in 3-space which are one-dimensional 'internally' even as they spread out through two- or three- dimensional regions in space (i.e. one moves along lines that connect points, which may be quite different distances apart, inducing the external scale-heterogeneity).

Finally there are, not surprisingly, though much less often taken as prototypical, also what we can call 'lamina', which are of co-dimension two in our three-dimensional space. These are sheet-like regions, with one

characteristic scale (the thickness of the sheet), such that points within the same sheet are more likely to interact with one another (either locally or across many scales of distance) than with points which are off or outside the thin sheet. Ideally, of course, lamina are internally 2-dimensional (the limit of zero 'thickness' of the sheets), just as networks or reticula are 1-dimensional. Real network connections (cables, rivers, roads) of course are characterized merely by a very large ratio between their length scales and any other spatial extension, and real lamina by the large area of the layer compared to its 'thickness'. There are again certainly natural laminar systems, such as the layered zones of lakes and seas by depth (and so also by salinity, temperature, pressure, and light levels), and perhaps also such phenomena as 'canopy ecologies' in rainforests, or layers of soils, etc. Artifactually, there is some isolation of connected underground levels of cities vs the surface and perhaps in some cases also of elevated levels.

In the case of lamina, there may be a single set of parallel laminar surfaces, in each of which interactions across both short and long distances within the sheet take precedence over interaction with nearby points outside the sheet. There may also of course be more complex 2-dimensional manifolds, the analogue of reticular grids, but I cannot off-hand think of ecological or artifactual examples, though there may be such phenomena in the complexly folded tissues of organisms (neural sheets?).

Extending the repertory of our prototypical images and metaphors of organizational complexity to include networks and lamina as well as spheres can help us to integrate both the 'levels' perspective and the 'scale heterogeneity' perspective in our analyses of self-organizing systems. My examples here have been based on the easily visualizable models of spatial scale relations, but of course dynamic temporal scales are equally if not more significant. One could generalize the embedded manifolds view of connectivity here from 3-dimensional space to an Einsteinian space of three spatial dimensions and a fourth one for time, but I will not explore this interesting approach here. It is sufficient to see that multi-scale and cross-scale temporal processes, as well as space-scale heterogeneity, are implied by the generalized models sketched in this section and elsewhere in the paper. We need many more detailed empirical studies to help us build up a repertory of means for representing how processes on different characteristic temporal as well as spatial scales (deriving from common dynamical scales) are intimately relevant to one another. Where we identify such cross-scale phenomena against the background of relative separation of levels, we may speak of 'downward causation', but I think it should be apparent from the arguments advanced here that this notion has an even wider generalization and signficance when we consider systems characterized by scale-heterogeneity.

 

Afterword

Throughout this chapter I have made cross-reference to other contributors to this volume. Most of the contributors met near Aarhus in 1997 to discuss their draft chapters and the unifying themes of the volume. I want to add here a few notes to place the perspectives of this chapter in the context of some of the other issues raised by the volume as a whole.

In their Introduction, Emmeche et al. offer a useful categorization of views on downward causation. I believe that my own approach most closely corresponds to their 'medium-strong' position rather than to their 'weak' version. I do believe, in agreement with Finnemann, and Bickard, that in the processes of emergent self-organization the very most fundamental laws of a system are produced and changed. The development of a system over time and through interaction within an environment can lead to dynamical possibilities in principle unpredictable from a knowledge of the system at any one time. (Self-organizing, open dynamical systems cannot in general be analyzed at single moments of time. They exist in a sense only over-time, and across many temporal scales; they move or die.)

This is a much stronger view than the 'preformationist' perspective, according to which all the possible futures of a system are fixed by its composition, i.e. from below. If emergent organization in the dynamics of a system can be represented by attractors of the dynamics, as Emmeche et al., Finnemann, Andersen, and many others have proposed, then my position is that in at least some forms of emergent organization for sufficiently complex systems the attractors themselves change as the systems' processes become entrained in interaction with still larger-scale processes. Along the historical developmental trajectory of such systems the very ground of dynamical possibility moves. This can happen, as chaos theorists have observed, because non-linear systems amplify 'noise' from levels below, turning it into information. I would only add that what kind of information it gets turned into also depends on the larger-scale processes of the environment within which the system is itself a constituent.

I also agree with Bickhard's response to arguments such as those summarized by Kim: a physics in which non-local fields or dynamic processes, rather than entities, are primary does not require that the behavioral possibilities of (process) constituents be independent of the organizational patterns in which they may be included. This is more obvious perhaps at the social level. People have radically different behavioral possibilities because they live in a complex technological culture: possibilities which were not available to our remote ancestors not simply for lack of the technologies and social institutions, but also because our species' bodies and brains have evolved, and our individual bodies and brains now develop, in the context of these larger ecosocial

systems and they now have different potentialities as well. In both Bickard's argument and mine, non-locality, or violation of the assumption of separable scales for constituent processes vs. aggregate processes is fundamental. A focus on entities, be they particles or organisms, makes scale-homogeneity seem more realistic than it actually is. 'Things' always fit neatly inside some sphere at some definite scale; fields and processes do not. Many arguments against 'downward causation' depend on the assumption of definite, separable scales. (So, of course, do many formulations of what 'downward causation' is.)

Such arguments also seem to depend on a synoptic perspective, that is, one that stands outside of time and ignores both history and change-in-progress. This is the temporal-scale analogue of non-localizability. Over time, the distinction between different scales which seem obvious at a single moment can become quite blurred (as in the example above of ecosystem effects on the evolution and development of brains, not to mention the role of changing brain capacities in altering ecosystems). Not only don't systems of the kind we are interested in really exist in single instants of time, the processes which constitute them are not confinable to characteristic time-scales.

I do still want to qualify my support for the medium-strong version of downward causation in two respects. First, like Patee and Moreno & Umerez, I do not believe that the metaphor or paradigm of cause-effect itself can be applied in an unrestricted way to the analysis of all phenomena. I believe it reaches the limits of its usefulness precisely (a) for systems of sufficient complexity that they must be treated as individuals with irreversible histories (and so for which there can be no question of same antecedents, same consequents) and (b) for cases where phenomena result from self-organization among many components and it is pointless to single out some of these as causal agents. I would rather say that the dynamics of constituent processes become entrained in the dynamics of larger-scale processes than that a larger-scale system simply 'causes' its constitutents to behave in novel ways.

Secondly, many arguments about downward causation assume a two-level model, whereas I believe that useful accounts must always consider at least three levels simultaneously. Processes at level L+1 represent selectional constraints on the possible ways processes at level L (in focus) can deploy the affordances they have by virtue of their constituents at level L-1. Historically, developmentally, phylogenetically, and probably cosmologically, we do not climb up from isolated quarks (or whatever) to complex organisms and ecosystems as the two-level, rung-to-rung model suggests. Instead, there is always already a higher level of organization (or at least of interaction; a quark soup, not just single quarks; ripples in the hyperdimensional continuum on many scales, not just the particle scale) and emergent self-organization produces more organized, more specified intermediate levels of order. We do not go from A to B to C to D ..., but from A-Z to A-L-Z to A-G-L-Q-Z .... In

this view what should be meant by 'causation', either upwards or downwards, may thus ultimately have more in common with notions such as material and formal cause than with the classical notion of efficient causation.

 

 

References

 

BAKHTIN, M. 1935. Discourse in the novel. In M. Holquist, Ed. The Dialogic Imagination (1981). Austin TX: University of Texas Press.

--- . 1953. Speech Genres and Other Late Essays. 1986 edition. Austin: University of Texas Press.

BATESON, G. 1972. Steps to an Ecology of Mind. New York: Ballantine.

BOHR, N. 1934/1961. Atomic Theory and the Description of Nature. London: Cambridge University Press. [reprinted 1987, Woodbridge, Conn: Ox Bow Press]

--- . 1958. The Philosophical Writings of Niels Bohr: Essays 1932-1957 on Atomic Physics and Human Knowledge. pp. 23-31. New York: Wiley. [reprinted 1987, Woodbridge, Conn: Ox Bow Press]

BOURDIEU, P. 1990. The Logic of Practice. Stanford CA: Stanford University Press.

--- . 1991. Language and Symbolic Power. Cambridge, MA: Harvard University Press.

EDELMAN, G. 1992. Bright Air, Brilliant Fire. New York: Basic Books.

GIBSON, J.J. 1979. The Ecological Approach to Visual Perception. Boston: Houghton-Mifflin.

GOODWIN, C. 1995. Seeing in depth. Social Studies of Science 25(2): 237-284.

GREIMAS, A. & COURTES, J. 1982. Semiotics and Language: An Analytical Dictionary. Bloomington, IN: Indiana University Press.

HASAN, R. 1992a. Meaning in sociolinguistic theory. In K. Bolton & H. Kwok, Eds., Sociolinguistics Today. London: Routledge.

--- . 1992b. Speech genre, semiotic mediation, and the development of higher mental functions. Language Sciences 14(4): 489-528.

HASEGAWA, A. 1985. Self-organization processes in continuous media. Advances in Physics 34(1): 1-42

HUTCHINS, E. 1995. Cognition in the Wild. Cambridge, MA: MIT Press.

KIRSHNER, D. & WHITSON, A., Eds. 1997. Situated Cognition: Social, Semiotic, and Psychological Perspectives. Hillsdale, NJ: Erlbaum.

KRISTEVA, J. 1980. Desire in Language. New York: Columbia University Press.

LATOUR, B. 1987. Science in Action. Cambridge, MA: Harvard University Press.

--- . 1990. "Drawing Things Together" in Lynch & Woolgar, 1990. [= 'Visualization and cognition: thinking with eyes and hands' Knowledge and Society, 6, 1-40. 1986.]

--- . 1993. We Have Never Been Modern. Cambridge, MA: Harvard University Press.

--- . 1996a. On interobjectivity. Mind, Culture, and Activity 3(4): 228-245.

--- . 1996b. Pursuing the discussion of interobjectivity. Mind, Culture, and Activity 3(4): 266-269.

LAVE, J. 1988. Cognition in Practice. Cambridge, UK: Cambridge University Press.

LAW, J., Ed., 1991. A Sociology of Monsters. London: Routledge.

LEMKE, J.L. 1984. Semiotics and Education. Monograph in Toronto Semiotic Circle Monographs Series, Victoria University, Toronto.

--- . 1985. "Ideology, intertextuality, and the notion of register" In J.D. Benson & W.S. Greaves, Eds., Systemic Perspectives on Discourse (pp.275-294). Norwood, NJ: Ablex Publishing.

--- . 1991. "Text Production and Dynamic Text Semantics." In E. Ventola, Ed. Functional and Systemic Linguistics: Approaches and Uses. [pp. 23-38]. Berlin: Mouton/deGruyter (Trends in Linguistics: Studies and Monographs 55).

--- . 1994. "Discourse, Dynamics, and Social Change." Cultural Dynamics 6(1): 243-275. [Special issue, Language as Cultural Dynamic, M.A.K. Halliday, Issue Editor].

--- . 1995a. Textual Politics: Discourse and Social Dynamics. London: Taylor & Francis.

--- . 1995b. "Literacy, Culture, and History." Review of The World on Paper by D.R. Olson. The Communication Review 1(2): 241-259.

--- . 1996. "Self-Organization and Psychological Theory." Review of L.B. Smith & E. Thelen, Eds., A Dynamic Systems Approach to Development. Theory & Psychology 6(2): 352-356.

--- . 1997. "Cognition, Context, and Learning: A Social Semiotic Perspective" in D. Kirshner and A. Whitson, Eds., Situated Cognition: Social, Semiotic, and Psychological Perspectives. Hillsdale, NJ: Erlbaum.

--- . in press, a. "Analysing Verbal Data: Principles, Methods, and Problems" in K. Tobin & B. Fraser, (Eds). International Handbook of Science Education. Kluwer Academic.

--- . in press, b. "Semantic Topography and Textual Meaning" in R. Stainton & J. DeVilliers, Eds., Communication and Linguistics. Toronto: Glendon College.

--- . in press, c. "Typological and topological meaning in diagnostic discourse." To appear in Discourse Processes.

LEONTIEV, A.N. 1978. Activity, consciousness, and personality. Englewood Cliffs, N.J. : Prentice-Hall.

LOVELOCK, J. 1989. The Ages of Gaia: A Biography of Our Living Earth. Oxford: Oxford University Press.

LUCY, J. 1992. Language Diversity and Thought . New York: Cambridge University Press.

LYNCH, M. & WOOLGAR, S., Eds. 1990. Representation in Scientific Practice. Cambridge, MA: MIT Press.

ODUM, HOWARD T. 1983. Systems Ecology. New York: John Wiley.

OLSON, D. R. 1994. The World on Paper. Cambridge & New York: Cambridge University Press.

PEIRCE, C.S. 1992. The Essential Peirce: Selected Philosophical Writings, Vol.1. Bloomington, IN (USA): Indiana University Press.

PRIGOGINE, I. 1961. Introduction to the Thermodynamics of Irreversible Processes. New York: Interscience.

--- . 1962. Non-equilibrium Statistical Mechanics. New York: Wiley.

--- . 1980. From Being to Becoming: Time and Complexity in the Physical Sciences. New York: W.H. Freeman.

PRIGOGINE, I. & STENGERS, I. 1984. Order out of Chaos. New York: Bantam.

SALTHE, S. 1985. Evolving Hierarchical Systems. New York: Columbia University Press.

--- . 1989. Self-organization in hierarchically structured systems. Systems Research 6(3): 199-208.

--- . 1993. Development and Evolution. Cambridge: MIT Press.

SAUSSURE, F. DE. 1915/1959. Course in General Linguistics. New York: McGraw-Hill.

SCHNEIDER, E. 1988. Thermodynamics, ecological succession, and natural selection. In B.H. Weber, D.J. Depew, and J.D. Smith, Eds., Entropy, Information, and Evolution. Cambridge, MA: MIT Press.

SILVERSTEIN, M. 1979. "Language structure and linguistic ideology." In P. Clyne, W. Hanks, & C. Hofbauer (Eds.), The Elements: A Parasession On Linguistic Units And Levels, pp.193-247. Chicago: Chicago Linguistic Society.

SMITH, L.B. & THELEN, E. 1993. A Dynamic Systems Approach to Development. Cambridge, MA: MIT Press.

THIBAULT, P. 1997. Re-Reading Saussure. London: Routledge.

THREADGOLD, T. & KRESS, G. 1988. Towards a social theory of genre. Southern Review, 21, 215-243.

VON UEXKULL, J. 1926. Theoretical biology. New York: Harcourt, Brace.

VYGOTSKY, L. 1963. Thought and Language. Cambridge, MA: MIT Press.

For "Closure: Emergent Organizations and their Dynamics"University of Ghent, Belgium; May 1999

 

Opening Up Closure: Semiotics Across Scales

Jay L. LemkeCity University of New York

jllbc@cunyvm.cuny.edu

Semiotics and Dynamics

I would like to offer what I believe is an interesting hypothesis about the relationship between semiotics and the dynamics of complex self-organizing systems. It assumes that the interesting complexity of such systems arises from the emergence of new levels of organization over their history. The fundamental proposal is that each new emergent level of organization in the dynamics of the system functions to re-organize variety on the level below as meaning for the level above. In this way, both the semiotic and the dynamical closure of system levels is re-opened to allow the development and evolution of greater complexity.

In order to clarify just what this proposal means, it will be necessary first to consider a number of concepts on which it is based. I will need to specify what I mean by levels of organization in a complex system, and how they relate to one another dynamically. For this I will rely mainly on the 3-level paradigm of Salthe (1985, 1989, 1993). Next, I will outline the kind of semiotic relationships that I believe can exist between levels. For this I will introduce a variant of Peirce's (1992, 1998) semiotics in which a basic distinction will be made between categorial meanings and meanings based on continuous variation. Finally, I will propose that emergent levels of organization tend to re-organize continuous variation at the level below as categorial information for the level above, and vice versa, perhaps in a hierarchy of alternating transformations between these two varieties of meaning.

It seems clear that biological systems are indeed multi-level systems of the kind to which this proposal should apply, and despite my very limited knowledge of contemporary theoretical biology, I will try to illustrate in a very rudimentary way how the proposal works in the case of the hierarchy of biological organization. I believe that examples from human sociocultural systems add further complications to the simple picture I will be sketching, and perhaps some of these also have precursors in simpler biological systems. I will mention some of these possibilities towards the end.

Hierarchies of Scale in Complex Systems

Certainly for biological systems, and probably for many others as well, the richness of their complexity derives in part from a strategy of organizing smaller units into larger ones, and these in turn into still larger ones, and so on. Such scale hierarchies are well recognized in modern biology (Salthe 1985, 1993). The parameter which most simply defines differences in scale, usually of a quantitative order of magnitude or more, may be taken to be the mass of a unit of organization on some level, the linear distance scale of strong correlations or interactions among the constituents of a unit, the energy scale of characteristic processes in which the units participate (typical amounts of energy exchanged in such processes), or the characteristic durational times of the cycles or processes which constitute the unit. As we move from level to level up the scale hierarchy, units get more massive, bigger, more energetic in aggregate (but with less energy used per interaction on the relevant scale), and slower in operation.

The dynamical relations among adjacent levels are what most concern us here. If we designate a level-in-focus, on which there is some emergent phenomenon constituting some units-of-interaction on a characteristic scale (level N), then in the 3-level paradigm of Salthe (1985, 1993; see Figure 1) we assume that units on level N are constituted by interactions at level (N-1) among the units at that lower level, but that of all the possible configurations which such interactions might produce at level N, only

those actually occur which are allowed by boundary conditions set at level (N+1). The logic of this dynamically is that the stability of a configuration of level (N-1) units at level N depends on the putative interactions among these new units at level N. Such higher-order interactions do not generally occur in an infinite and empty vacuum, but rather in some context (a container, a medium, an environment) some of whose properties (e.g. temperature, pressure, ambient energy flows) are specified by still larger-scale dynamics (e.g. solar insolation, atmospheric pressure, salinity). Only those configurations at level N will be even meta-stable on the appropriate time-scale which are consistent with the constraints imposed from level (N+1). The properties of units and interactions at level (N-1) are constitutive for level N phenomena; those at level (N+1) are constraining for level N phenomena. (Salthe 1985 refers to 'initiating conditions' from N-1 and 'boundary conditions' from N+1.)

Figure 1. [See also Appendix B]

This is still a synchronic, or steady-state view of hierachically organized systems. More truly dynamically, the model assumes that new levels of organization always emerge between previously existing levels. New complexity arises in systems because the new level N re-organizes the relationship between level (N-1) and level (N+1). Level N units and their interactions now mediate between the levels above and below: not all variety at level (N-1) remains available for re-organization at level (N+1), there is a filtering performed by level N. Conversely, we may say that level (N+1) is buffered against variations at level (N-1) by the stabilizing mediations at level N. This principle is closely related to the model of Ehresmann and VanBremeersch (1996, 1997) in which each level of organization can be realized by a variety of combinations at the level

below; thus changes at that lower level do not result in qualitative differences at level N, and may not produce any effects at level (N+1). This is also qualititatively similar to Thom's principle of structural stability (1975).

At the same time that the emergence of stable units and processes at level N reduces the flow of information from level (N-1) to level (N+1), thus in one sense simplifying the dynamics of the system (there are fewer possible dynamical combinations allowed), and so making it more specified (cf. Salthe 1985, 1993 on the 'specification hierarchy' and see Lemke 1994), the emergent level now also means that a new kind of information must be given in the description of the system; in this sense it has become more 'complicated'. We need new descriptive categories to talk about the units and phenomena of level N. Insofar as the emergence of level N may occur developmentally only following the prior emergences needed to produce its immediate precursor organization (the original level N-1 units, level N+1 conditions, and their dynamic relations), the system has also now reached a new degree of 'logical depth' (cf. Collier 1999, Collier & Hooker n.d.).

Furthermore, the new organization of the system now presents us with new options for its further development, specification, or evolution. Once some particular units and phenomena at level N have emerged, it is now possible for still newer levels to be interpolated between (N+1) and N, and between N and (N-1). Each new level of organization augments the evolutive potential of the system for increasing its organizational richness still further.

Before leaving the issue of dynamical scale, I want to add one more point that will be relevant later on. For typical biological systems, such as cells or multicellular organisms, all of the scale parameters listed above (mass, size, energy, time, etc.) tend to shift in step with one another and it does not much matter, beyond analytical convenience and the availability of data, which we use. But this is not always the case. An important critique of the application of systems theory to human communities and their technological infrastructures has been made by Latour (1987, 1996), who notes that our usual systems theories assume a particular topology: we assume that units closer in space are always more likely to interact than more remote units are. This leads to assuming a 'spherical' topology for system: we envision its levels of organization as nested spheres of larger and larger spatial scale. But it is perfectly possible for two distant units to interact more intensively or more frequently than two nearby units, if the distant units are connected by a stable channel of communication, while the nearby units are not. This leads to a 'network' topology in which units on the same network interact more than units that may be nearby in space but are 'off' the network. People on the internet but geographically distant may communicate much more frequently and influence one another's actions more than happens with their close geographical neighbors who are not on the net (or not part of a specific social network). In an ecosystem, one locus and another downstream from it may be more tightly coupled than the same locus and a nearer one that is isolated from the stream. In an aquatic or arboreal system with stratified layers, two distant sites within the same layer may be more likely to interact than either is with spatially closer sites in an adjacent layer -- at least in respect of the flow of certain kinds of information, nutrients, pollutants, etc.

In these more general cases, what happens to the neat notion of scale levels? I believe that it can be shown that what matters in the general case is not specifically spatial

scale, but time scale (Lemke, in preparation). This is so to the extent that what makes a "level" in a system is its dynamical functions in relation to other levels, and insofar as the key dynamical function I am proposing is semiotic reorganization, what matters is the informational or communicational coherence of a level (and its relations to adjacent levels). If we think in terms of subnetworks within a network topology, then the way in which scale levels can be defined is in terms of the adiabatic principle, which insures that levels are relatively insulated from one another if the timescales of their characteristic and constitutive processes are sufficiently different that they cannot exchange significant amounts of energy on each other's relevant timescales. Such fastnets and slownets can each be internally communicatively coherent while remaining relatively communicatively isolated from one another, as required by the 3-level paradigm. They can only be constitutive (N-1 to N) or constraining (N+1 to N) in relation to one another, and not directly interacting. This timescale perspective promises to be especially valuable in analyzing sociocultural ecosystems, or ecosocial systems (Lemke 1994, 1995), but is also fundamental to many other kinds of analysis (e.g. entropy and information measures may also be timescale-specific).

Topological vs. Typological Semiosis

My basic proposal is going to be that each new emergent level serves to reorganize one type of semiotic information from the level below it as another type for the level above it. What are these two basic types of semiosis? I will follow the basic model of C.S. Peirce (1998), with a few specializations of terminology, to explicate these types. Semiosis is a process of meaning-making. It is a process of construing some material entity or phenomenon as a sign, rather than simply interacting with it energetically. In Peirce's terms, the entity or phenomenon is called the representamen (R), and what we take it to be a sign of is called the object (X). But Peirce wisely recognized that no R directly points us to a corresponding X; there is work of interpretation to be done, there are principles or codes by which this interpreting is done, and so there must be, in my terms, a system of interpretance (SI). A sign is only a sign for some SI; R is a sign of X only for some SI (and not necessarily for all SI's; for some the same R may signify a different X, or the same X be signified by a different R, nor need these relations always be one-to-one).

The first step in defining a role for semiotics in the dynamics of multi-level systems is to map these defining elements of semiosis onto the 3-level paradigm of scale organization, as in Figure 1. Units or phenomena on level N are representamina, R, of object-states, X, of the interactions of units at level N-1 which dynamically constitute the phenomena at level N, for processes or structures at level (N+1), which form the system of interpretance, SI, with respect to which correspondences between R's and X's are defined and computed. Note that this implies that the SI always has both a slower timescale, and usually a more global spatial-extensional scale, than the phenomena which it interprets.

It is useful here to note the precise sense in which semiotic interpretation differs from simple physical interaction. When an organism encounters some photons or some molecules of a particular chemical species, there is a physical interaction of these entities with the organism's sensors on the corresponding scale (e.g. molecular). That is interaction. But if the organism as a whole, mediating and buffering the molecular-scale consequences of this interaction through the dynamics of other higher-scale levels of its

internal organization, reacts in a way that is adaptive to the presence of food, or predators, in its environment, then it is reacting on a higher scale level than the purely physical, initial molecular-scale interaction, and it is thus acting as a system of interpretance, as if it were interpreting the encounter with the molecules or photons as a sign of the presence of food or predators (in Peirce's terms, an 'indexical' sign). The organism reacts to the molecules not just as molecules, but also as telltales of and cues for response at a higher-scale level.

There is nothing here which non-material; semiosis is a material process, but one in which typically there is a translation or re-interpretation of information from one scale level to another. This introduces the possibility of many-to-many mappings of information as we move to macroscopic scales of organization, while fundamental interactions, at the molecular or quantum scale, tend to be one-to-one. Microscale interactions are deterministic because there are very few dynamical degrees of freedom. Photons and electrons have no individuality; most small-molecule reactions have unique chemical outcomes, there is simple deterministic cause-and-effect. But when effects at the microscale are re-organized at higher scales, it becomes possible for the same molecular encounter to produce a positive tropism in one species but a negative tropism in another. The same response may be elicited by different stimuli in the same organism, the same stimulus, under different conditions of context at higher scales, can produce different responses in the same organism. Semiosis does the work of interpretation in large part by integrating contextual factors across scales (Lemke, in press-a). Now we no longer have simple deterministic causality; many causal inputs on many different scales combine in ways that lead to unpredictable and emergent behaviors, some of which prove adaptive. The extension of the anthropomorphic metaphors of 'interpretation' to simpler biological and even non-biological material systems are carefully considered in Anderson et al. (1984) and in the work of Hoffmeyer and Emmeche (1991; see also Hoffmeyer 1997).

For those who may be more familiar with Peirce's view of semiosis, it may be useful to clarify a few points. Others may wish to skip this paragraph and the next. While a sign, or more properly a representamen, R, does not determine its object, X, directly, it does, for Peirce determine its interpretant, which is another sign whose object is the relation between the initial R and its X. My own view, and probably Peirce's, is of course that R does not cause or create the interpretant: this work is done by the SI. But what enables R to function as a sign of X for the SI is that, given how the SI interprets R's, this R will be interpreted as the appropriate interpretant sign, and so as a sign of the appropriate X. It is not a causal or physical 'determination', but a logical or semiotic one. Now in my view the interpretant sign is always materially instantiated at a scale level above that of the initial R. According to Peirce, the interpretant sign normally becomes in its turn a representamen that determines some still further interpretant, and so what is for Peirce the basis of an indefinite chain of semioses is for me the basis of an indefinite (i.e. open-ended) hierarchy of scale levels of systems of interpretance. This is not the place to consider whether this 'unlimited semiosis' ultimately leads to a difficulty with the number of scale levels available, but it does pose an interesting question about the scale of the object X.

For Peirce, all along the chain, the interpretants are giving information about the same object, though they may lead us to see it in different ways or make different inferences about it. In the dynamical hierarchy view I am constructing here, each higher-scale SI at

level (N+1), which interprets representamina on the level N below it, construes doubly: the material reality which is being construed by the SI as a representamen is some particular pattern of correlations of material interactions among constituents at level (N-1), but at the same time the interpretant or meaning (at level N+1) of this pattern-as-sign corresponds to an object at its own (N+1) level. So in one sense the object X of the original representamen is materially grounded at level (N-1), but in the second sense, the object X corresponding to the interpretant is some phenomenon that has material relevance at level (N+1). When I smell something, in one sense my smell-perception is a perception of the molecular interactions in the olfactory bulb, but it is semiotically (conceptually) interpreted as an index of some macroscopic food or waste or predator. A sign is meaningful for an SI because it is a sign of something on the scale of the SI. We smell the molecules of our enemy as our enemy; it is, normally, the enemy, not the molecules that are dangerous, relevant, and meaningful. Are these 'the same' object X? In the model being developed here we would need to be bit more careful in our answer than perhaps was necessary for Peirce.

In order to continue my larger argument, I need now to distinguish two broad classes of semiosis: (a) those cases in which the features of representamen that are criterial for some SI to interpret it as a sign of some X may vary continuously, so that quantitative differences of degree in a feature of R normally lead to differences of degree or kind in the interpretant, vs. (b) those in which all representamina are classified by the SI into a discrete spectrum of types, and each R-type is interpreted as a distinct X. I will call the first case 'topological semiosis'; it is a generalization of the notion of analogue signaling. By 'topological' here I really mean to invoke the topology of the continuum of the real numbers: it is continuous variation, quantitative differences of degree that matter. I will call the second case 'typological semiosis'; it is a generalization of the principle of digital signaling. By and large most of classical (Saussurean) semiotics restricts itself to typological semiosis, but this is insufficient for the analysis of general system dynamics.

Figure 2. [See also Appendix B]

The complications here arise from the fact that in principle we may map continuous variation in X onto continuous variation in R (and vice versa), the usual 'topological' case, or discrete variants of X onto discrete variants of R (and vice versa), the usual 'typological case' (Table 1), but we may also have mixed modes of semiosis in which

the continuous is mapped onto the discrete and vice versa (Table 2). The mixed modes are relatively rare in human cultural conventions for symbolism and representation, but I believe they are fundamental to the inter-level relations of dynamical systems.

Thus typically, as in linguistics, what is represented (X, a semantic category) and how it is represented (R, a string of words) are both discrete. Semantic categories have a meaning potential which is defined by their contrasts with other discrete categories. A noun may be Singular or Plural, but it cannot have some continuously specifiable degree in between Singular and Plural. A verb may be Present tense or Past tense (or some one of a closed and finite list of other discrete possibilities), but there are no continuous degrees between Present and Past in natural languages. The word(s) that represent these differences are likewise discrete: a word (R) must be interpretable either as 'man' or as 'men'; pronouncing a vowel somewhere on the acoustic continuum between them does not make a new word in the English language, and cannot represent some semantic degree in between Singular and Plural. This is not how natural languages work. They are almost completely typological in their mode of semiosis. (There are some exceptions, in the case of intonation, for example.)

Table 1. Typological vs. Topological Semiosis

Type-X represented as:

spoken word written word mathematical symbol chemical species

Quantitative-X represented as:

size, shape, position color spectrum visual intensity pitch, loudness

 Iconic:scale models, maps

 Indexical :voltmeter, thermometer

 Symbolic:cartesian graph scientific visualization

We adopt this principle also for our mathematical and scientific symbol systems. A chemical element may be Carbon or Nitrogen; there is no continuum of elements (so far as we know) between them, and we represent them by discrete symbols: there is "C" and there is "N", we do not adopt the real number line to represent atomic species, as we do, for example, for atomic weights (even though they do not vary continuously, they may have almost any arbitrary ratio to one another). The variables in our equations are either 'x' or 'y' and we do not allow continuous variation between variable x and variable y (though we do allow for linear combinations in order to unify typological and topological logics, which is in some sense the basic historical function of mathematics, or at least of algebra, algebraic geometry, and real analysis; see Lemke, in press-b).

But we do indeed need and use the topological mode of semiosis as well. In the pure case, this means representing a continuously variable X by a continuously variable R, or equivalently, interpreting continuously variable R as continuously variable X. Thus we typically represent continuously variable dynamical parameters of a system by real numbers, or by positions on a line or in a 2- or 3-dimensional space (e.g. Cartesian graphs). We could, and with the new methods of computerized scientific visualization increasingly do, represent them by the continuous visible color spectrum, or by degrees of brightness, or by acoustic pitch or loudness. Scale models and maps represent continuously varying shapes by continuously variable shapes (iconic signs); a thermometer or barometer represents by the continuously variable height of a column the continuously variable temperature or pressure of the surrounding system (indexical signs). In Peirce's classification scheme, the arbitrary conventions of the Cartesian graph illustrate topological symbolic signs. Topological semiosis is at least as important and general as typological semiosis in the representation of nature and its dynamical systems.

As an aside, we might note that our scientific accounts of natural phenomena are conceived in terms of a mixture of typological concepts, in verbal language and the discrete variables of our theories and equations, and in the discrete types of entities that we recognize and visualize, together with topological representations of continuously variable features of systems, where that continuous variation matters to some phenomenon of interest. Mathematics is the semiotic bridge by which we bind together our primitive perception and representation of continuous variation (size, shape, spatial position, brightness, loudness, color, temperature, etc.) and our more categorial conceptual languages. Mathematics is in essence and historically an extension and specialization of the semantic categories of natural language to deal with continuously variable phenomena, or what may as well be such relative to the observer's scale.

If 'information' is, in Bateson's (1972) famous phrase 'a difference that makes a difference', then it is clear that both difference of kind, and difference of degree can make a difference. Each can be the basis of semiosis, each can be interpreted as representing a difference in the object X for which some representamen R stands for a particular system of interpretance, SI.

What happens in the mixed-mode case where we represent quantitatively variable phenomena (X) by discrete or typological representamina R? We certainly do this, as the examples in Table 2 show.

Table 2. Mixed-mode Semiosis

Quant-X as Type-R :

waveform as phoneme

painting as description

ratios as fractions

Type-X as Quant-R

words as sonogram

numeral as bitmap

semantic category as fuzzy set

functions as algebraic expressions

conformations as ligand classes

event-type as probability?

 

In speech, there is continuous acoustical variation in sound, but for an organism that has learned a particular language, only certain distinctive feature differences count or 'make a difference' to which vowels or consonants (phonemes) and so which words are heard. When we write speech down with letters of an alphabet, we transcribe continuous variation into strings of discrete variants. We preserve the purely linguistic information (which words are said), but we lose much other information (whose voice was speaking, with what regional accent, in what degree of agitation). The SI classifies or 'types' what we hear into a finite number of discrete and mutually contrasting equivalence classes. To do so it has to have learned the language, or at least learned its 'phonology'; this learning is a process on a much longer time scale, requiring even in its momentary application to perceived sounds many more neuronal units and interlinked networks, on a much larger system scale, than the simple registering of acoustic degrees of pitch and intensity. We hear the same word even when we do not hear exactly the same sounds, e.g. from different speakers, when someone has a cold or is nervous, etc. The phoneme-analysis level of the neurological speech perception system buffers word-scale recognition from much shorter timescale, non-criterial fluctuations in perceived sound. It filters out the 'noise' that is not criterial for categorial differences between phonemes or words. The mapping of the sound-, phoneme-, and phonological system -responsive networks of the brain onto the 3-level paradigm should be obvious, as should its fit to both the dynamical scale hierarchy levels and their semiotic functions.

Consider another example at a more molecular level. When proteins fold up into their complex 3-dimensional spatial conformations, then at the scale of an amino acid what we see is discrete typological variation: each constituent unit is either this amino acid or that one, there is no continuous variation among kinds of amino acids. At the scale of the protein as a whole, interacting with something else at the same scale, what matters is the continuously variable distances and angles of its local conformational shape and electromagnetic fields. But at the still higher scale of a membrane with binding sites composed of many interlocking proteins, the complex spatial configuration of a ligand matters only insofar as it does or does not occupy a site and produce some triggering effect. The membrane reads only discrete ligand classes in many cases, and is blind to the details of the conformation, so long as they are within certain parameters. Many of our successful medical drugs are simply 'imposters' which fool the membrane, or some other complex larger-scale structure, because they are indistinguishable as members of the ligand equivalence class defined by the membrane, which is thus operatingas a higher-scale system-of-interpretance (or part of such a system).

In addition to writing systems, other human semiotic conventions also operate this particular mixed mode. When we represent the rational numbers as fractions, we are representing something that varies quasi-continuously in terms of discrete pairs (integer numerator and denominator). When we represent continuous functions by algebraic equations (e.g. polynomials), the representamina are discrete 'x' and 'y' just as if they were words in a verbal sentence (and indeed we can read algebra as if it were a sentence

of English with only slight peculiarities of grammar, because of the historical origins of algebraic notation).

What of the converse mixed mode? What happens when we represent discrete phenomena by continuously variable representamina? This happens for instance when we look at a sentence of English as a 'sonogram' or acoustical graph of energy across a range of frequencies, displayed longitudinally in time on an oscilloscope or continuous feed tracing roll. You can learn to read sonograms as if they were a form of writing. It is quite difficult because there is so much 'extraneous' information represented in the sonogram, information that does not matter to deciding what words were said. The neurological networks that have already been trained to filter sound for phoneme classes do not transfer over very readily to screening these visual patterns. In fact you very quickly realize that in 'hearing words' we are often 'hearing' sound cues that are not actually there acoustically; as the sound stream is matched to the most likely or only possible word sequence, higher level networks are activated that correspond to whole words or phrases, even if parts of those words, or sometimes whole words within phrases, are not instrumentally detectable in the sonogram! This is somewhat of a linguistic analogue to Gestalt pattern completion phenomena for visual perception.

Using sonograms as a system for visually representing spoken language is not nearly as efficient as alphabetic, or even ideographic writing systems because it tries to represent a typological system of semiosis topologically. It requires the creation of a whole new system of interpretance rather than piggy-backing on an existing one. Nonetheless there may be some advantages in some cases to such a procedure. The new 'fuzzy engineering' represents semantic categories of verbally stated criteria for good machine functioning by continuously varying functions for degrees of membership in a category (the 'membership function' for fuzzy sets and fuzzy logic). By doing so it can 'smooth out' transitions in behavior and more closely approximate arbitrary functions.

Figure 3: Trans-organization across modes

Level N-1 Topology to Level N Typology

Continuum dynamics to eigenvalue typology Bifurcations, attractors, basins

Threshhold effects

Topological variety to equivalence classes

Fuzzy sets to sharp sets ??

Level N-1 Typology to Level N Topology

Discrete items averaged to net densities, concentrations

Discrete units organized as polymers, lattices, networks

Discrete lattice or network dynamics organized to coherent macro-phenomena

("propagation") and molar property effects ("elasticity")

Discrete events organized as continuous action (neural firings --> smooth motor actions)

Considering both the logic of the 3-level paradigm, in terms of how semiotic functions are mapped onto dynamical scale levels, and many examples such as those just given of the reorganization of continuous variation into discrete variants (Figure 3, upper), and of discrete variants into continuous variation (Figure 3, lower), has led me to what seems at least heuristically an interesting conjecture:

The Principle of Alternation:

Each new, emergent intermediate level N in a complex, hierarchical, self-organizing system functions semiotically to re-organize the continous quantitative (topological) variety of units and interactions at level (N-1) as discrete, categorial (typological) meaning for level (N+1), and/or to re-organize the discrete, categorial (typological) variety of level (N-1) as continuously variable (topological) meaning for level (N+1).

In each case, level (N+1) functions as the system of interpretance which construes entities and phenomena at level N as signs of microstates of the system at level (N-1). By extension, where these level (N-1) states correspond to the effects of interaction with the environment at level (N-1), higher levels of the system respond to them as signs at level N of phenomena in the environment at level (N+1), which may have only a very indirect causal-material relationship to the actual interactions at level (N-1) or none at all. It is not my purpose here to discuss an interactional model of how semiosis mediates learning; for an interesting effort in this direction, see Bickhard & Turveen (1995). My concern here is with the logic of the Principle of Alternation itself.

Dynamics and the Principle of Alternation

The basic mapping of semiotic functions onto organizational scales in Figure 1 has a dynamical implication, if we interpret it in terms of the evolution or development of the system:

Principle of Emergence:

A new level in the scale hierarchy of dynamical organization emerges if and only if a new level in the hierarchy of semiotic interpretance emerges.

This is in some sense a logical precondition for the Principle of Alternation. The exact connection between the two becomes clear if we ask ourselves whether the semiotic relationship between adjacent levels of the dynamical scale hierarchy could be a simple mapping of continuous variation at the level below to continuous variation at the next level up? Or of discrete variants at the level below to discrete variants at the next level? This is logically possible, but would we then consider that there was any point in saying that a qualitatively new level had emerged? And if it had, what would its functional advantages be? In a mapping of continuous variation onto continuous variation, there is

very little room for novelty or innovation; there is only re-description. Similarly for mapping one discrete set onto another; what else is this but re-naming, especially if it is one-to-one? As Ehresmann and vanBremeersch (1996, 1997) argue in the case of a logical model of hierarchical multi-scale systems, the novelty of new levels arises in part because each higher level has many possible realizations at lower levels. It is only a many-to-one mapping that provides for classification, and filtering and buffering (from level N-1 to level N), as the 3-level dynamical scale model expects. The 3-level model of course also expects that as we go up one more level, a categorial element (at level N) can be interpreted (at level N+1) to have many possible meanings, or system responses, depending on contextual constraints from 'elsewhere' (at level N) that are integrated by the higher-scale (more global, longer timescale, N+1) level.

I believe that this dynamical logic appears to us users of human categorization as an alternation between topological and typological semiotic relationships of adjacent levels, taken in (possibly overlapping) groups of three. This is clearest in the cases where quantitative variability is reduced to discrete categorial variants. We know of many applicable mathematical models in which continuum dynamics produces discrete states, when subject to higher-scale boundary constraints: discrete spectrum eigenvalue solutions, bifurcations, discrete attractors and their (classificatory) basins, threshhold effects of all kinds. What is a bit less obvious is how the complementary half of the cycle of alternation proceeds. How do systems interpret discrete variants as continuous variation?

The answer in all cases is the same as for the continuum-to-discrete part of the cycle: by going up one level in scale. We have already seen one example of this: protein polymerization. At the small-molecule scale, the protein and its interactions in vitro are defined by the discrete typology of the constituent amino acids (level N-1). But when we consider the protein molecule as a whole, and interactions at a higher scale level (level N) that depend on, say large-molecule to large-molecule interactions, then it is the continuously variable conformational shape of the folded protein that matters. A shape which is co-determined both by the amino acid sequence (from level N-1), and by contextual-environmental constraints (level N+1) of the overall global cell chemistry (which determines, for example, the cytoplasmic pH, temperature, or similar global conditions). We should not be surprised that what at the more micro- scale look like discrete units, appears from a more macroscale perspective as continuous variation. This is the basic molecular-to-molar logic of chemistry. At the pauci-molecular scale (Halling 1989, Kawade 1996) reaction pathways depend on discrete, non-stochastic interactions of specific molecular species; but when we proceed to larger scales, such as global cellular chemistry, then we are closer to the regime of concentration-dependent effects where the Law of Mass Action applies. Concentration-dependent effects and chemical gradients are large-spacescale, long-timescale averages over discrete molecular interactions.

If we consider neurocortical activity in the brain, even in a simple model in which neural 'firing' is all or nothing (discrete variants), as we move up in scale we eventually find that there are global coherent phenomena that average over many individual 'firings' to produce the alpha and other well known EEG rhythms of continuous variation. Karl Pribram's (1991) famous 'hologram hypothesis' also posits that functionally meaningful patterns are construed more globally across neurological activity. We also know that individual firings of nerves that activate bundles of muscle

fibers are globally coordinated at a higher scale (and a longer timescale) to produce smooth motor action of an entire muscle or muscle group.

Thus discrete items may be averaged to net densities and concentrations, discrete units organized into polymers and lattices (which have global coherent effects such as elastic propagation modes, which are again quasi-continuously variable phenomena), and discrete events globally coordinated to produce smooth, continuous higher-scale actions.

What the Principle of Alternation proposes is that the transformations of discrete to continuous and continuous to discrete alternate as we move from level to level of the dynamical hierarchy, and that in doing so they represent a semiotic transformation of the information content of lower levels as signs for higher levels, allowing many-to-one classifications and one-to-many context-dependent reinterpretations. A scale of dynamical processes in a system, at which such transformations occur, meets the logical conditions for novelty that define for us a genuinely emergent level of organization.

Table 3 (see also Appendix A) illustrates a possible sequence of such alternations from level to level, though clearly our, or at least my, knowledge of all the intermediate scales is too limited to present it as more than a suggestion of the plausibility, or at least the heuristic value, of looking at the dynamics of multi-scale systems from this perspective.

Table 3. The Principle of Alternation -- Examples

Quantum variety (typo) organized as molecular charge distributions (topo)

Biomolecule conformations (topo) organized as ligand class information (typo) by larger-scale membrane polymers

Pauci-molecular reaction pathways (typo) organized as molar concentration-dependent effects (topo) at global cell-chemistry scale

Molar chemistry (topo) organized as neuro-transmitter threshold effects (typo): "firing"

Firings in neural nets (typo) organized as coherent cortical effects (topo): "brainwaves" "holograms"

Cortical dynamics (topo) organized as limit cycles (typo): "percepts" "phonemes"

Neuronal attractor effects (typo) organized as smooth motor behavior (topo): "drawing" "gesticulating" "enunciating"

Smooth motor behavior (topo) organized as visual and verbal signs (typo): "gestures" "words" in ecosocial supersystem as meta-system of interpretance

See also Appendix A for discussion.

 

Alternation and the Re-opening of Closure

Let us return, finally, to the theme of closure. In what sense are self-organizing systems closed? And what is the relevance of closure to the evolution of complexity? We must first distinguish several quite different meanings of 'closure' with respect to such systems. There is material closure, which would mean that there is neither matter nor energy flow across the boundary of the system. There is autocatalytic closure in the sense that some web of interdependent processes is self-regenerating. There is informational closure in the sense that all information critical to the system's behavior is available internally. There is semiotic closure, which entails that in some sense the system's dynamics depends on exhaustive sets of classificatory alternatives. And there is the well-known semantic closure thesis of Pattee (e.g. 1995), which posits that the semantics of classificatory symbols completes the dynamical description of such systems by specifying initial conditions on general dynamical laws.

Pattee's thesis has much in common with the argument being presented here. He emphasizes what I have been calling typological semiosis, and speaks of the type categories as 'symbols' which have both a local material structural instantiation and a function in relation to a more global system organization that 'interprets' them. His concern is not explicitly with levels of scale in this process, but he does follow von Neumann's arguments to posit "multiple-level descriptions when we need to relate structure to function." It is not clear if these are only distinct logical levels, or also scale levels, as proposed here. In any case, my principal concern will be rather with the narower issue of semiotic closure, as above, rather than with Pattee's very general notion of semantic closure, which is already an integral part of my proposals. (I might however differ with his conclusion that only natural selection can explain the symbolic dimension; across levels there is an intimate dialectic between physiological process and already-semanticized structures, leading to new symbolic functions that depend as much on material self-organization as on selectional accumulation of stored information.)

Clearly the kinds of self-organizing, dissipative multi-level structures in which we are interested are not materially closed, for they are always energetically parasitic on larger-scale energy flows, small negentropic back-loops riding piggy-back on the greater downhill degradation of order to disorder. Here, too, the issue is a matter of scale: the scale -- in mass, energy, spatial extension, and most critically in time -- of self-organization is always vastly less than the scale of the dissipative energy flows which support them. As Prigogine (1961, 1962) noted, it is both the flows through the system, and their adjoint larger-scale external constraints that together allow smaller-scale entropy reductions far-from-equilibrium. Moreover, where the dynamics is truly complex, there are many 'entropies', many modes of dissipation, and only some of them decrease in correspondence with the spontaneous emergence of order in the correlations among only some of the system's dynamical variables (Hasegawa 1985). We focus our attention on these particular variables, because their correlations have meaning for us. It is only dynamical correlations that we define to be phenomena, and about which we make our science; we privilege order because we see ourselves as the children of order, and our science is, not surprisingly, always also about us.

Of course it is also true that systems and networks are definable as units of analysis, on each scale, because their internal interdependencies and communications are more intensive than their overall dynamic dependency on interactions with their environments. The latter are the precondition of their being, but the former define their

specific characteristics. In the usual 'spherical' topology of systems (as above) there may be a closed boundary separating system from environment (e.g. a membrane), and this may be a precondition for the development of delicate smaller-scale internal dependencies, buffered from the larger-scale fluctuations outside (Hoffmeyer 1997b). But in more general network topologies the buffering may be provided by the adiabatic principle itself (external fluctuations are too slow on the scale of internal processes to affect them significantly on their own characteristic timescales) or by channels of connectivity (phonelines, watersheds) which strongly link system elements internally (at high speed, or with high matter or energy fluxes) but without either enclosing the 'system' or blocking flows from outside -- they simply promote flows inside and let the adiabatic principle do the rest. Thus systems with 'network' rather than 'spherical' topologies can overlap and interpenetrate in real space; they can also interact, on a slower scale, at their 'centers' (or anywhere), rather than only at their margins or boundaries. This is a very different sort of material 'closure' once again.

The notion of autocatalytic closure (Kauffman 1993, Hoffmeyer 1997b) is a material-informational condition on self-organizing systems of sufficient complexity, and a plausible one. The system is in effect capable of synthesizing its own constituents down to some minimum level of organization (atoms, small molecules), but at and below that scale it is dependent on material flows from the environment, as also on environmentally maintained energy gradients and conditions sufficient for disposal of waste heat and other toxic by-products of the synthesizing processes. This notion has close similarity to that of semiotic closure, but in a multi-scale levels-of-organization perspective we cannot say that a system is informationally closed on autocatalytic grounds: lower levels provide informational input in the sense of smaller-scale types that are functional in the system at the scale-in-focus, and higher levels provide information in the sense that it is functional integration into these levels that determines the conditions of interaction of such systems with others on their own scale. We have perhaps too easy a habit of imagining our systems as operating solo in some much less-structured environment: one protocell in the primordial soup. But the more typical case is one in which there are other cells in the soup, and beyond the soup higher-order structures (whether tissue-like or ecological) which depend upon and constrain interactions between these units. From the cell's (organism's) point of view (internalist perspective, cf. Matsuno 1989), it is interaction with other cells (organisms) that provides the information input that we macro-observers (externalist perspective) ascribe to the next higher level of multi-cellular (ecological) organization.

In order to better formulate issues of closure across multiple levels of organization we may wish to ask in what sense our systems of interest may be semiotically closed? It would seem that self-organization does imply a degree of informational and dynamical autonomy, but to what degree and in what sense autonomous? The epigenetic principle of equifinality implicates our homeostatic, or better homeorhetic (Waddington, 1957) tendencies: larger-scale, longer-term processes and the structures they give rise to are buffered by intermediate scales of organization from fluctuations at lower levels … within some limits of tolerance defined by scale ratios (larger-scale changes in the gradients that sustain the system energetically are not buffered against, only energetically smaller-scale fluctuations, and not all of those, but only the ones already implicated in the evolution of the intermediate levels of organization). So long as this buffering is successful, the kinds of order and even the specific units, processes, and structures on each level of organization vary within narrow parameters: the

informational description of these types is closed. There is nothing new under the normal sun.

Semiotically, each higher level is characterized by its own exhaustive paradigms of types, and dynamically the inter-level relations (for adjacent levels) have come to be (ontogenetically, as phylogenetically) such that normal fluctuations at lower levels do not matter because they do not, by construction, alter the 'structural stability' (in the sense of Thom, 1975) of the structures they constitute at the higher level. So far as the system-of-interpretance at the higher level is concerned, these are differences that do not make a difference; they have no information value, they are merely yet more points in phase space that all lead onto the same attractors. It is the attractors and their basins that define the higher level dynamics, closing it semiotically at the same time they presuppose its openness materially and energetically.

But semiotic closure as a typological notion can never be the whole story. We all have two eyes, but they are never the same distance apart, never the same exact shape or size. All faces look much alike at some level of typological classification, but their quantitative 'topological' differences allow us to distinguish them as individuals and to form, at a new intermediate level of interpretance, Gestalt patterns (again types) of recognition from many quantitative features, and then classify these patterns yet again. In development, the type-specific features are equifinal, but there is still plenty of room for quantitative individuation. At levels of organization where only typological difference matters, and for levels for which this is true, we can speak of semiotic closure within a level. But if the Principle of Alternation is a useful guide, then across semiotic triples of levels, there is always somewhere a lack of topological-semiotic closure, and it is this very source of potentially meaningful open variation which is reorganized at some higher level again into a new typological-semiotic closure.

We should not be surprised that in self-organizing systems quantitative variability is organized at a higher-scale level into qualitative invariants. We tend, however, in looking at cross-level relations only two at a time, rather than three at a time, to focus only on how typological closure arises from metric openness. When we consider multi-scale systems across many triples-of-scale, we also see the alternating inverse process by which larger-scale aggregations of many lower-scale types begin to appear once again as quasi-continuous distributions: the raw material for yet new orders of order to emerge at new intermediate-scale levels.

We also make more sense of closure if we look at the development (in the individual instance) or evolution (of the type-class) of a hierarchy of orders of closure at various scales of organization. According to the 3-level paradigm, new emergent levels of organization come-to-be between always pre-existing scales (primordially between smallest units and global flows/constraints that permit dissipative structure to emerge initially, and then again and again, always within supporting prior structure and residual free energy). This is a dynamical hypothesis, based on semiotic motivation. It says something about how additional levels of organization get added to an initial multi-scale dynamical system (perhaps including the primordial case of a non-semiotic two-level system). The typological order at some level, with respect to the SI level above, always has some residual quantitative variability that initially is 'ignored' by the level above. But it can (and presumably must, though 'under what conditions?' remains a key question) happen that certain correlational patterns within this residual quantitative

variability become significant for the level above. This does not add new types to the original lower level, it remains semiotically 'closed'; but the correlational patterns in quantitative relations among these units now constitute a new intermediate level just insofar as they do come to matter to the level above.

These three levels now constitute a new semiotic triad, and all of them are changed by the emergent order 'in the middle' (this emergent order is really relational across all three levels; we just tend to focus on the middle level because it is the level at which the new types are defined). The top level (of these three) is changed because it has new informational input from below, it is now selectively sensitive to new kinds of pattern or order among its constituents that it was previously insensitive to. The middle level is of course entirely new. The lower level is also changed because it is now subject to new constraints from above: in the presence of the correlational patterns that define the next higher level, there are fewer degrees of freedom accessed by the lower level. Why do these particular correlational patterns, among all the possible ones, come to be selected out as signficant at a higher scale? We can say because they are 'adaptive' at the higher scale, but what we mean now is that they are the lower-scale signs of something that has reality, that makes a difference on larger scales, particularly on longer timescales, than the ones on which they themselves exist. It may be that they are the system's way of feeling ahead into its future, or at least into the future with respect to the timescale of the level in focus. One can say, with a pointlike-present view of time, that they aid 'anticipation' at some level (Rosen 1985, Salthe 1997). Or equally we can say that they are the means by which the temporal coherence, the organization and integration of processes across different timescales, is achieved level by level (or triple by triple).

We are also saying not just that these particular correlational patterns are selected for because they are adaptive in the above sense, but also that the available patterns that can be selected for, arise in large part from the attractors of the material dynamics of the constituents of the original lower level, subject to the constraints of the original higher level. Not only are they not random, but there are very few available possibilities, and different lineages can be expected to work small variations with the same available raw material. Selection becomes fine-tuning, but the instrument is largely wrought by its own self-organizing dynamics, a dynamics that in the case of biological systems has a very great evolutionary and developmental 'logical depth' (Collins & Hooker 1998).

In the story I find myself telling here (which is only a hypothetical one, of course), 'logical depth' corresponds to the unavailability of short-cuts in the ontogeny of a highly evolved, or generally a multi-level, self-organizing system: new meanings of new patterns of quantitative correlation among types on a lower scale cannot emerge until those types, along with their possible residual quantitative variations and relationships, have previously emerged. That emergence is never predictable the first time, for it is a function of unique conditions, unique to the individuation of some particular system, within the developmental trajectory of its kind before the emergence. And once it becomes somehow conserved (inevitably as a function of the longer-timescale persistence, or recreation, of those originally unique -- or at least not yet known to be persistent -- conditions) as part of the future trajectories of such systems, still newer dynamical possibilities come into being whose very terms of definition could not have been formulated one stage back, because the types whose residual quantitative variation is the basis for the new possibilities did not yet either exist or necessarily need to come to be. Some very different raw material might have become available, and some

different selection adaptively shaped. Viewed as a process of computation, the results at each stage must be computed before it is possible to even define the algorithm for computation of the next stage. There are no computational shortcuts, no way to predict the instance-specific individuation component in development, only the type-specific recapitulative component, and so no way to predict evolutionary futures insofar as either long-timescale constraints may change, or the details of future evolution (and so of the late type-specific stages of future developmental trajectories) depend on shorter-timescale processes that provide the raw material to be shaped by even persistent long-timescale constraints.

Heterarchy and Heterochrony in Ecosocial Networks

That the evolutionary principle just described applies equally to human social history (i.e. to the evolution of technology, mores, and more generally 'ecosocial systems', Lemke 1994, 1995), and indeed to the history of science (or 'technoscience', Latour 1987, 1993), should be obvious. The implications, which deny the fantasies of social engineering which justify our current technocratic legitimations of power, are understandably refused by all those who hope, or will, that we can make history just as we please. We can, should, and must 'push' and 'choose' on the timescales it is given us to act within, but without the illusion that we can foresee the consequences beyond the next unpredictable emergence.

There are moreover a number of additional complications in the case of multi-scale systems, or more generally multi-scale networks (Latour 1996) of interdependent semiotic-dynamical processes in which material entities with semiotic value are produced on one scale, but are also interpretable on other scales. The topology of networks seems the most appropriate one for such systems, and, as I would argue in this case (Lemke, in preparation), timescales provide the most relevant parameter for defining cross-scale semiotic-informational relationships. A semiotic artifact may persist over very long periods of time, and itself be the product of large-scale social-institutional processes over long timescales, and may accrue and accumulate (relative to some larger-scale, longer-timescale cultural system of interpretation) many meanings, but nonetheless it may also be interpreted by human-scale actors on relatively short timescales and play a role in determining our short-timescale actions. In this way semiotic mediation can occur between processes on very different timescales, between which purely dynamical connections are forbidden by the adiabatic principle. This introduces new semiotic connectivities, with definite dynamical consequences, into the networks. What is often called 'cultural evolution' has its specific dynamical significance through these 'heterochronic' connections, perturbing a neat dynamical hierarchy toward a less well-understood semiotic-dynamical 'heterarchy'. I believe that this is the key phenomenon that must be considered in making a dynamical theory of ecosocial systems. It seems reasonable to suppose that it has precedents in biosemiotics, though whether or not DNA itself could be considered such a heterochronic material-semiotic 'artifact' is not entirely clear to me.

Regardless of these important complications in the case of human sociocultural, semiotic ecosystems, we can still, if the hypothetical story I have been telling in this paper is at all a useful guide, look to the (multi-dimensional) residual quantitative variability within and between our naturalized conceptual categories as a potential resource for opening up the closure of meanings they impose on us. Every sharp

categorial distinction that we can analyze into quantitative 'fuzziness', on as many dimensions as possible; every finer-scale specification which we can backtrace to a less-differentiated, more pluripotent 'vagueness' ; every conflation that maps categories onto one another one-to-one that we can prise apart into a combinatorial matrix of relative frequencies of association … enlarges our meaning-space and multiplies our possibilities for imagination and so for action … making the future ever more unpredictable and rich in new orders of organization to be temporarily superimposed on the gradual and far larger-scale universal degradation of order, just as we ourselves are.

References

Anderson, M., Deely, J., Krampen, M., Ransdell, J., Sebeok, T.A., & von Uexkull, T. 1984. "A semiotic perspective on the sciences." Semiotica 52: 7-47.

Bateson, G. 1972. Steps to an Ecology of Mind. New York: Ballantine.

Bickard, M. and Terveen, L. 1995. Foundational Issues in Artificial Intelligence and Cognitive Science: Impasse and Solution. New York: Elsevier.

Collier, J.D. 1999. "Causation is the transfer of information" In Howard Sankey (ed.),Causation, Natural Laws and Explanation. Dordrecht: Kluwer.

Collier, J.D. and Hooker, C.A. n.d. "Complexly organized dynamical systems." (manuscript, submitted to Open Systems and Information Dynamics).

Ehresmann, A. C. & Vanbremeersch, J.-P., 1997. Information processing and symmetry-breaking in memory evolutive systems. BioSystems 43, 25-40. -- and see under vanBremeersch, below.

Halling, P. 1989. "Do the laws of chemistry apply to living cells?" Trends in Biochemical Sciences 14: 317-318.

Hasegawa, A. 1985. Self-organization processes in continuous media. Advances in Physics 34(1): 1-42

Hoffmeyer, J. 1997a. Biosemiotics: Towards a New Synthesis in Biology. European Journal for Semiotic Studies, Vol. 9 No. 2, pp 355-376.

Hoffmeyer, J. 1997b. Surfaces Inside Surfaces: On the Origin of Agency and Life. Paper presented at the 9th International Conference on Systems Research, Informatics and Cybernetics, Baden-Baden, August 18-23, 1997.

Hoffmeyer, J. and Emmeche, C. 1991. Code-Duality and the Semiotics of Nature. In Myrdene Anderson and Floyd Merrell (eds.):On Semiotic Modeling. Mouton de Gruyter, Berlin and New York. pp. 117-166.

Kauffman, S. 1993. The Origins of Order. Oxford: Oxford University Press.

Kawade, Y. 1996. "Molecular biosemiotics" Semiotica 111: 195-215.

Latour, B. 1987. Science in Action. Cambridge, MA: Harvard University Press.

--- . 1993. We Have Never Been Modern. Cambridge, MA: Harvard University Press.

--- . 1996. On interobjectivity. Mind, Culture, and Activity 3(4): 228-245.

Lemke, J.L. 1994. "Discourse, Dynamics, and Social Change." Cultural Dynamics 6(1): 243-275.

--- . 1995a. Textual Politics: Discourse and Social Dynamics. London: Taylor & Francis.

--- . In press, a. "Material Sign Processes and Ecosocial Organization." To appear in P.B. Andersen, C. Emmeche, and N.O. Finnemann-Nielsen, Eds. Downward Causation: Self-organization in Biology, Psychology, and Society.

--- . In press, b. "Meaning in the Middle: Measure, Picture, Gesture, Sign, and Word". To appear in Anderson, M., Cifarelli, V., Saenz-Ludlow, A., & Vile, A., (Eds.), Semiotics Perspectives on Mathematics Education. Erlbaum.

--- . In preparation. "Across the Scales of Time: Multi-level Organization in Activity Networks and Ecosocial Systems." For Mind, Culture, and Activity (Erlbaum).

Matsuno, K. 1989. Proto-Biology. Boca Raton, FL: CRC Press.

Pattee, H. 1995. "Evolving self-reference: matter, symbols, and semantic closure." Communication and Cognition - Artificial Intelligence 12: 9-27.

Peirce, C.S. 1992. The Essential Peirce: Selected Philosophical Writings, Vol.1. Bloomington, IN (USA): Indiana University Press.

Peirce, C.S. 1998. The Essential Peirce: Selected Philosophical Writings, Vol.2. Bloomington, IN (USA): Indiana University Press.

Pribram, Karl H. 1991. Brain And Perception: Holonomy and Structure in Figural Processing. Hillsdale, N.J.: Lawrence Erlbaum Associates.

Prigogine, I. 1961. Introduction to the Thermodynamics of Irreversible Processes. New York: Interscience.

--- . 1962. Non-equilibrium Statistical Mechanics. New York: Wiley.

Rosen, R. 1985. Anticipatory Systems. New York: Pergamon.

Salthe, S. 1985. Evolving Hierarchical Systems. New York: Columbia University Press.

--- . 1989. Self-organization in hierarchically structured systems. Systems Research 6(3): 199-208.

--- . 1993. Development and Evolution. Cambridge: MIT Press.

--- . 1997. "Modeling Anticipative Systems" In Advances in Modeling of Anticipative Systems, G.E. Lasker, D.Dubois & B. Teiling (eds). Int. Inst. Adv. Stud. in Syst. Res. and Cybernetics.

Thom, R. 1975. Structural Stability and Morphogenesis. Reading, MA: W.A. Benjamin.

VanBremeersch, J.-P., Chandler, J., & Ehresmann, A. 1996. "Are interactions between different timescales a characteristic of complexity?" ECHO Symposium, Amiens 1996.

Waddington, C.H. 1957. The Strategy of the Genes. London: Allen & Unwin.

Appendix A.

Examples of Alternation in the Biological Hierarchy: Discussion

Quantum variety (typo) organized as molecular charge distributions (topo)

Quantum units at the scale of atoms or ions represent discrete, categorial, typological variety. An atom is either a Carbon type or a Nitrogen type or some other element; there are no intermediate degrees of Carbon-ness, no normal atomic or ionic species with a fractional number of protons. As we look at a biological molecule on a scale somewhat larger than the atomic scale, the variety available to be organized is the set of discrete differences from one atomic species to the next. We may increase this repertory somewhat by considering the different bonding arrangements and allowing that a nitrogen amidst one set of atomic neighbors does not behave as exactly the same intra-molecular species as a nitrogen in some other atomic environment, but still the set of possible types is finite and discrete. However, as we begin to consider the molecule as a whole as an emergent unit of organization at a new level, and we ask what properties of the molecule over longer many-atom stretches are relevant to how other molecules react to it, then we see that it is the electrical charge distributions in space, which have a quasi-continuous representation, which matter. Molecular organization represents a re-organization of the discrete, typological variety of atomic species into quasi-continuous spatial distributions of electrical charge, as a function of the interactions among the species in various combinations; and this new toplogical variation in electrical charge is what conveys information to the next higher level of organization: intermolecular interactions.

We see this most strikingly in the case of proteins, where the discrete units are compounded from atoms to repeating amino acid units (again slightly different in different neighboring AA environments, but still a finite set of types), but what matters in terms of the action of the protein as a whole in the larger-scale environment is a spatial conformation, the folded-chain, which is a collective effect and presents us with emergent properties on a larger spatial scale, which are properties that matter to a still larger scale, as below. Typological variety is emergently reorganized at a larger spatial-material scale as topological variation that is meaningful for phenomena at a still higher scale.

Biomolecule conformations (topo) organized as ligand class information (typo) by larger-scale membrane polymers

What is this next scale? The topological variety, regarded as information, represented by the conformations of macromolecules (e.g. folded proteins) and their associated (interactively functional) spatial charge distributions matter to a still higher scale, the chemistry of the cell, only by way of, for example, their interactions with intermediate-scale cellular structures such as membranes and interior reticula. The membranes, of many kinds in the cell, have in common that they respond to biomolecules as ligands, that is, as potential binders to membrane sites. These sites are formed by the interaction of the membrane-constituent molecules and represent, in their own spatial conformations and active charge distributions, an emergent level above that of the individual biomolecular species. These active sites define equivalence classes for other biomolecules, ligand classes, effectively those which will or will not bind to the site and produce some effects. In some cases binding may be partial, and some ligands may produce more pronounced effects (say the opening or closing of a membrane pore), but what has

happened here is that, so far as the next still-higher level is concerned, all that matters about a biomolecule is which ligand class it belongs to relative to this membrane site.

The topological variety of the folded protein of the ligand, or of the site itself, does not matter in its quantitative detail, but only via these equivalence classes and discrete effects. Topological variety has been reorganized at a new emergent level as typological meaning for still higher levels. (Note that obviously there are still some quantitative effects that are matters of degree, but the new organizational level as such fits what is expected from the principle of alternation.)

Pauci-molecular reaction pathways (typo) organized as molar concentration-dependent effects (topo) at global cell-chemistry scale

Suppose we now move up again in scale, from a view in which we see individual molecules interacting to one in which we see only statistical distributions and average concentrations. In between, there is the interesting regime of what is coming to be known as pauci-molecular chemistry, where the assumptions of the law of mass action and macro-chemistry are not met. There may be further emergent levels of relevant organization at these intermediate scales, but they are not well understood yet. Nonetheless, it is quite clear that in this regime we are in transition from typological phenomena in which there are again quite discrete reaction pathways dependent on local conditions, to a higher level of global or gross cellular chemistry at which the law of mass action is a better or very good approximation. Global level cell chemistry has its own emergent properties, such as overall pH and other average concentrations, and these variables are now topological in nature, because to good approximation they are continuously variable. (See Halling 1989, Kawade 1996).

Molar chemistry (topo) organized as neuro-transmitter threshold effects: "firing" (typo)

The molar chemistry of cells, representing topological variety, can again be reorganized at still higher levels into typological variety, as in the well-known case of nerve cells that ‘fire’; certain quantitative thresholds are exceeded leading to global chain reactions throughout the cell, and we know that still higher level brain processes depend on configurations and sequences of this now typological variety: cells that do or do not cascade or discharge. Whole multi-cell synaptic sequences are built, which are, like molecular species built from atomic ones, again of discrete identifiable types (each recurrent pathway is a type in this sense).

Firings in neural nets (typo) organized as coherent cortical effects (topo): "brainwaves" "holograms"

But how do these synaptic cascade pathways matter to larger brain processes? in part at least they matter by way of global, coherent electrical excitation of the brain, such as the alpha wave patterns and others of similar kind. These wave patterns, while themselves discrete types, carry information in a topological form: continuously variable amplitudes. We should also note here that it appears that there are many other such global, or at least ‘volume’ effects of neurotransmitter concentrations, such that neurons may not actually ‘fire’, but only carry slow-wave changing potentials, influenced by peptide concentrations and modulating the graded release of neurotransmitters that affect many neurons in the local volume. Here too we find higher-scale topological effects of lower-scale typological variation, and vice versa. Another hypothesis of like sort is Pribram's "holograms" or holonomic functional units (1991) which in many brain systems appear to be organized non-locally across brain structures and their electrochemical activity.

Cortical dynamics (topo) organized as limit cycles (typo): "percepts" "phonemes"

It is at least possible in some models of brain functioning that topological, meso-scale brain dynamics, in the form of propagating waves of electrical activity, whose effects matter through degrees of intensity, whether of chemical concentrations or of electical polarizations, interact at a still higher scale of brain activity to produce emergent levels corresponding to elementary "percepts" or to "phonemes", which are frequently or regularly typological in their informational variety. This presumably happens through the

emergence of new attractors in the dynamics of the meso-scale functioning, each attractor in effect classifying its basin into a type.

Neuronal attractor effects (typo) organized as smooth motor behavior (topo): "drawing" "gesticulating" "enunciating"

Clearly there is an intimate interdependence between perceptual and motor functional elements in such a scheme, and we can again see a transformation or re-organization in the production of smooth motor behavior, which functions in terms of its topological characteristics (timing and coordination, gross and fine movement in space), which are in turn emergent from the discrete, typological elements that correspond to dynamical attractors in the neuro-muscular system. Ennervations and innervations of particular nerve elements and muscle fiber bundles (typological and discrete) emerge as overall continuous motion in space (topological, characterized by continuous degree and change). This is surely a miracle of emergent organization at a very high level (slower processes, on larger spatial-extensional and matter-energy scales).

Smooth motor behavior (topo) organized as visual and verbal signs (typo): "gestures" "words" in an ecosocial supersystem functioning as the meta-system of interpretance

Finally, I will end here with the last step of the link from physics to language and human social semiotics that I promised, namely the smooth motor actions (topological) are re-organized by learned processes of organisms in communities to be produced and interpreted as signs, such as word-utterances and gesture-productions, which are classic instances of typological signs. The SI here is not just the organism, but the organism in a community, and not just a community of other persons, but an ecosocial system that includes all the relevant nonhuman agents or actants as well (e.g. written texts). The timescales here are not just those of cognition, but those of language- and culture- learning, and indeed implicating the next higher scale: historical change in the social meaning systems for interpreting words and gestures.

 

For Language as Cultural Dynamic (Special issue of Cultural Dynamics, M.A.K. Halliday, Issue Editor)

DISCOURSE, DYNAMICS, AND SOCIAL CHANGE

 J.L. Lemke

City University of New York

Brooklyn College

Brooklyn, New York 11210 USA

 

The Problem of Cultural Dynamics

Which aspects of social and cultural change are in principle predictable and which are not? How can we usefully model the dynamics of such complex systems as human communities? What is the role of discourse, and of culturally and historically specific semiotic formations generally, in co-determining the processes of social change? How are discursive, semiotic practices and material, ecosystem processes inextricably linked in the dynamics of social systems?

In what follows I will attempt to sketch the broad outlines of a theoretical framework within which the answers to such questions may be sought, and I will suggest some tentative solutions to these general problems of cultural dynamics. I believe that the basic tools needed for these tasks have recently been developed in other disciplines, needing only critical re-interpretation to be applicable to cultural dynamics. In a brief exposition such as this, many details must be omitted, but they already are or soon will be available elsewhere (see citations below).

Sophisticated readers will be well aware of the hazards of theory-making in this field. It is all too easy to be seduced by single principles, or by partial perspectives, including those congenial to our own cultural biases or ideological class interests. It is notoriously difficult to bridge between microsocial and macrosocial approaches, or between the materialist and idealist traditions. Above all, it is difficult to make a truly dynamic account of social and cultural systems, since most of our traditional concepts (individuals, institutions, cultures, societies, languages, discourses) are formulated in essentially static terms.

There are more fundamental limitations on our ability to model systems of which we are ourselves a part. As observers and theorists we are limited by scale: we exist for mere decades, while the systems we seek to model exist for centuries; we can observe only small regions of space at any one time, while social systems extend over nations and continents. We change quickly: maturing, aging, dying; many cultural processes occur so slowly that they may not seem to us to be occuring at all. We are also limited by position: we are members of some cultures and societies and not others; we speak some languages, dialects, registers, and discourses but not others; we are socially positioned observers, with only a single gender, age, class, or status, commanding a limited range of viewpoints even within the social and culture groups of which we are members.

On the other hand, as evolutionary and social products of the systems we study, we are pre-adapted to model them (in strictly limited ways) as a condition of our own survival (cf. Rosen 1985 on `anticipatory systems’). As members of social groups we can participate in co-operative enterprises in which multiple observers cover larger areas, command a greater multiplicity of (not always easily shared) social and cultural viewpoints, and (partially) interpret the (incomplete) records left by our historical predecessors. It is as futile to imagine that a single organism will completely model its own ecosystem, or a single individual its own society, as that a single cell will model an organism, or a single molecule a cell. But a social system, a culture, might well construct and maintain a model of itself, necessarily incomplete perhaps (cf. G? 1931/1962), but possibly adequate for certain purposes.

Human culture has already constructed the key tool needed to make such a model of itself: a linguistic system which is, semiotically, its own meta-system. Any natural language may serve as its own meta-language, in which its own grammar, and more importantly a theory of how its speakers make meanings with that grammar (Halliday’s grammatics, in press-a, and this volume), may be articulated. A community deploys the resources of its language in discourses, social formations that define a particular way of using language to make a particular kind of meaning, and those discourses (including the discourse of cultural dynamics) may be employed to model of at least some processes of the whole cultural system which includes them. Discourses are made by the social and cultural interactions of many actual, individual speakers over a period of

time: interactions which are also simultaneously processes in the living material, ecological system which is their human community.

 

Social Semiotics and Cultural Dynamics

The theory of cultural discourses as social semiotic formations (Halliday 1978, 1988, in press-b; Lemke 1988a, 1989, 1990a, 1990b; Threadgold 1986, 1989; Hodge & Kress 1988; Thibault 1986, 1989a, 1989b) has arisen from the study of the semantics of texts inspired by Bakhtin’s social linguistics (Bakhtin 1929/1986, 1935/1981, 1953/1986) and Halliday’s functional semantics (Halliday 1985). It provides a model for a general theory of social formations, which is a key element in the theoretical synthesis we need.

Social semiotics begins by disputing the primacy of the sign and the exclusive emphasis on sign systems in formalist semiotics (e.g. Eco 1976). Instead it gives priority to the signifying act as instance, and to social signifying practices as regular, repeatable, recognizable types. Social semiotics offers the view that socially meaningful doings constitute cultures (social semiotic systems): that cultures are systems of interlinking, socially meaningful practices by which we make sense to and of others, not merely in explicit communication, but through all forms of socially meaningful action (speaking, drawing, dressing, cooking, building, fighting, etc.). Sign systems are abstractions from such practices (e.g. linguistic signs from speech), changing as social practices change.

Sign systems are semiotic resource systems; they enable us to make meaningful actions (including utterances) by deploying these resources in recognizable, mostly habitual (and marginally creative) ways. The habitual ways in which we deploy them are identifiable as semiotic formations: the regular and repeatable, recognizably meaningful, culturally and historically specific patterns of co-deployment of semiotic resources in a community. A literary genre of a period is a semiotic formation; so also is an architectural style and type of building, a religious ritual, a typical holiday meal, the making of a particular type of costume. All these formations are defined in terms of the regular patterning of actions, of social meaningful practices, that members of a community are engaged in when producing them.

Discourse formations are social semiotic formations in which the deployment of linguistic resources is essential to the social meaning of the result (though other actional semiotic resources may also be deployed as part of the formation, as with gesture in speech, graphics with writing, etc. Cf. Lemke 1987). The linguistic (semantic and grammatical) resources specific to a particular discourse formation form a register of the language (a specific distribution of the probabilities of deploying any meaning alternative the language provides; see Gregory 1967, Ure & Ellis 1974, Halliday 1978). A particular type of Weather Report, for example, would be a specific discourse formation, deploying a portion of the register of meteorology, and doing so according to a schema of organization and sequence which is often called a genre (Hasan 1985, Martin 1985, Ventola 1987, Lemke 1988b, 1990a). There are speech genres, and genres of both literary and non-literary writing; and there are also, more generally, action genres, which need not involve language at all, though their enactment may be guided by use of a discourse formation, as when we `talk ourselves through’ a complex performance or an activity requiring difficult, context-dependent choices.

Semiotic formations provide an intermediate level of conceptual analysis between the microsocial (utterances, texts, particular acts and events) and the macrosocial (dialects, institutions, classes, ideologies), but more importantly, they formulate the scale from microsocial to macrosocial in terms of actions (social practices) and patterns of relations of actions (cultural formations) and not in terms of entities and aggregations of entities (individuals: corporate groups: societies). This is an essentially cultural view: social systems are systems of doings, not of beings as such. They are systems of interrelated cultural practices, not systems of socially interacting individuals as such. The ultimate theoretical constituents of a social system are not interacting dyads, not even individual members, but individual social and cultural practices. Social `individuals’ must be theoretically re-constructed (e.g. Lemke 1988c); they are no longer `givens’ in a semiotic social theory.

The link between formations and macrosocial structural relations and dynamics is provided by another cultural notion: organized heterogeneity (cf. Wallace 1970 on `organization of diversity’). Social and cultural systems are not homogeneous, they exhibit an essential internal diversity: their subsystems present alternatives, complementary or conflicting (e.g. male vs. female practices, class conflicts, age-grade diversity, etc.). In the case of discourse formations (and more generally the language varieties and semantic orientations associated with differentiating social variables, cf. Bernstein 1981, Hasan 1986, Hasan & Cloran 1990), Bakhtin (1935) labelled this phenomenon heteroglossia. In every community there co-exist different regional and social group dialects, different historical usages, different modes of speaking associated with interest groups, age-groups, genders, ideological points-of-view, etc. Social semiotics identifies, with Bakhtin, both the ideational and the value-orientational relations among these different social voices (Lemke 1988a, 1989, 1990b, in press; Thibault 1989). Each subcommunity constructs a different reality by the views it formulates in language on any matter, and it constructs its views always and only from a particular social position of interests and values vis-a-vis other possible or actual views.

It is our own culture’s ideology that fuses the ideational dimension of language’s semantic resources (and of semiotic resources generally, cf. Lemke 1989, 1990a), with the real, the natural, the universal, the immutable, the absolute, while disjoining this from a contrasting fusion of the social-orientational dimension with the relative, the conventional, the cultural. Semiotic resources in themselves, the potential for cultural meanings, independently combine both the `social-cultural’ and the `natural’ with both `making it be’ (`action’) and `telling what is’ (`representation’ `knowledge’); cf. Halliday (this volume). Contrary to the ideology of our culture, what an analysis of the resources it deploys shows is that the `natural’ is just as `conventional’ as the social, the `real’ just as much a product of action as the `cultural’. We make both the social and the natural, or better: the natural and the cultural make each other and are one, not two. The natural has all the properties of the cultural: it is mutable, conventional, and relative to the orientation of the observer. Semiotically (i.e. epistemologically, and probably historically as well) the natural is derivative from the cultural, a specialization later made to seem an opposition.

What applies to the discourses of diverse and often conflicting subcommunities applies equally to all their social practices and formations: there is in general a system of heteropraxia, of specific relations of alliance, opposition, etc. among their ways of doing, each with respect to the others, of which the system of heteroglossia is one very

important part. No text can be read, no action interpreted, without taking into account the aspect of its meaning that derives from the existence or possibility of alternative and conflicting ways of saying and doing within the same total community. And no utterance or action escapes making meanings, anticipated or not, in these terms.

Semiotic formations are relatively stable elements in the flux of day-to-day social action; they insure the minimal short-term predictability necessary for social coherence. Formations may be represented in terms of their constituent actions as selections from sets of alternatives with contrasting meanings, each selected alternative implying something about the structure and sequence of action as well as the specific acts to be performed. This is a straightforward generalization of the paradigm-and-realization model of language used in Systemic Linguistics (Halliday 1976, Fawcett 1980). They may also be represented in some cases structurally (Hasan 1989, Ventola 1987) according to a syntagm-and-realization model, or according to a mixed approach appropriate to the kind of formation being described (cf. the representation of discourse formations by thematic patterns diagrams in Lemke 1983, 1988a, 1990a).

Since semiotic formations are co-deployments of resources that form systems of (paradigmatic) semiotic alternatives (often from different semiotic systems, e.g. language and gesture or picture, cf. Lemke 1987), and since they are in turn characteristic of and constitutive of divergent subcommunities (heteroglossia, heteropraxia), it is most general to represent them in terms of the conditional probabilities for the co-occurence of various practices in various contexts, according to the subcommunity, and indeed the culture as a whole. This can be done within the general relational-contextual model of meaning employed in social semiotics.

Any action or process (or simultaneous or sequential combination of such) is socially meaningful only in relation to other alternative actions or processes (and combinations) that might have occurred in its place. The specific meaning is interpretable only in relation to the set of socially relevant contexts which are constructed for the purposes of that interpretation. Such contexts are generally analyzable into syntagmatic contexts (events before and after), paradigmatic contexts (alternative events), and indexical contexts (situationally co-occuring events); see Lemke (1990a). All of these contextualization relations may be formally represented as meta-redundancy relations (Lemke 1984: 35-39; see also Halliday in press-a), which specify the conditional probabilities for co-occurence of various alternatives in various contexts, but hierarchically, so that higher-level alternatives (e.g. higher vs. lower social class) co-occur with entire probability distributions linking, say, semantic types of utterances to situational uses (cf. Hasan 1990), and not with particular acts or situation-types separately. Thus to be a member of a social-class subculture is not to use only some semantic strategies available in the language, or even to use them with a certain distinctive probability, but to combine them differently with the demands of situation from what a member of another class might do.

The irreducible formal hierarchy of contextualization in social meaning reflects the dynamical hierarchy of emergent levels of organization in human social-material systems, as we will see below. This connection is already implicated in the social semiotic view of meaning. The semiotic systems of a community are abstractions of the resources in actual use in that community. The semiotic formations present their habitual patterns of co-deployment, and the metaredundancy relations summarize their

mutual co-occurence distributions with respect to each other. All these analytic forms are abstractions of types from tokens; they all depend entirely on the moment-to-moment happenings in the community. It is instances (events, acts, occurrences, performances) which are primary.

No inertia or active constraining force is attributed to semiotic systems like language, formations such as genres, or metaredundancy relations. If these abstractions from practice behave as if they were dynamical systems (e.g. by persisting, by changing coherently, by evolving over time), it is because the instances from which they are derived are themselves simultaneously elements of genuine, material dynamical systems. If formal semiotic relations exhibit an irreducible hierarchical organization, it is because the dynamical systems which enact them do so. Semiotic systems and formations, and thus culture, do not have an autonomous dynamics of their own, but rather a complex dialectical interdependence with the material dynamics of social communities.

Although social semiotics avoids the dead-end of idealist reification, it still only allows for, and does not yet provide a true dynamics, which must equally avoid the trap of a naive materialism. Every scientific discourse of material processes and relations is still a discourse, a semiotic formation which stands inside some culture and not outside culture altogether. Every instance, even a nuclear detonation, is only meaningful insofar as we can construe it as an instance of some type(s) already (or nascently) provided by our meaning system (system of semiotic practices). But meaning systems, and cultures, change; what was not meaningful before can become meaningful, and this process depends critically on a dialectic between material and semiotic dynamics within a total (material-cum-social) system. We need to understand the dynamics of the larger eco-physical systems in which cultures are embedded and from which the energy of their dynamics derives. Human social communities are material ecosystems. Even though physics, chemistry, and eco-biology are themselves simply cultural discourses just as linguistics and social semiotics are, these two different orders of discourse construe two different kinds of relations among events and processes, and it is the connection between those different kinds of relations that is the key to modeling cultural dynamics.

 

The Dynamics of Complex Systems

We may regard a social practice as a `type’, as a semiotic, cultural abstraction, but every particular, actual instance of that social practice is enacted by some material process in a complex physical, chemical, biological, ecological system. Every action thus enters into two systems of relations, for which our culture has two different sorts of descriptive discourses. As an instance of a social practice, it enters into relations of meaning with other social practices. These are semiotic relations. As a physical event, it enters into relations of energy, matter, and information exchange with other events. These are material relations. Every instance of a social practice is simultaneously also an instance of some material process. Every system of social practices, linked in semiotic formations according to their meaning relations (cf. the metaredundancy relations), is also a system of material processes linked by physical, chemical, and ecological relations.

When we build a building, we quarry stone or cast concrete, we construct doors and walls and windows, we build floors and stairs and shafts, we place ducts and vents, all in accordance with a system of cultural practices that defines for us an architectural style, a desirable design according to historically and culturally specific discourses of what buildings and rooms should be like, how comfort and privacy should be provided for, what spaces are monumental, which public, which private, how size and shape and light should co-occur with use and function: in short, with the architectural semiotics of our culture.

But in doing so we also assemble masses with densities, weights, compressional and tensile strengths, electrical and thermal conductivities; we arrange flows of water and air according to principles of hydrodynamics, flows of heat by principles of thermodynamics, and allow for material flows of people and goods. We may provide a system of communications capable of handling certain rates of information transfer, a supply of energy that may flow from solar panels on the roof to heating channels in the walls, or from underground generators through cables to electrical connections in every room. These things we do also according to discourses of our community, those regarding science and engineering, but in all these doings, our actions belong simultaneously to the order of the semiotic and the order of the material (itself articulated, modeled, through the semiotic). The relations we construct are simultaneously semiotic relations in a cultural meaning system and material relations in an eco-physical system.

Every meaningful social practice can be enacted only through some material processes, and the semiotic formations that link practices through their cultural meanings for us inevitably couple material processes which have other actual and potential eco-physical relations in a system of an entirely different order of discourse. In this way, whether in obvious cases such as the construction of cities or the clear-cutting of rainforest, or in less evident ones such as the publishing of books, the imprisonment of offenders, the selection of mates, or the setting of wages and prices, cultural linkages of social practices into semiotic formations produce eco-physical linkages of material processes. And reciprocally, the linkages of material processes on which the eco-physical being of the community depends, which indeed are the eco-physical being of the community, form the ground of all possible and actual change in the relations of semiotic practices.

The full implications of this intimate dialectic of practices and processes, semiotic-discursive and material-ecophysical relations cannot be fully appreciated without the realization that both aspects of a total ecosocial system are hierarchically organized at many different scales through complex couplings of processes which feedback on one another to produce entirely surprising, emergent phenomena. In the dynamics of complex, tightly coupled systems with strong multiple feedback loops, even small regularities can produce surprising global effects. Semiotic formations, which slightly bias the linkages of material processes according to their semiotic meanings for a human culture, are essential elements in the material dynamics of human communities, and this material-semiotic coupling is reciprocal. There cannot be two systems here, changing according to separate laws, relatively independent of one another. There can be only one unitary ecosocial system, material and semiotic, with a single unified dynamics, described under two aspects, by two different sorts of culture-specific discourses.

The unity of ecosocial systems is somewhat hidden from view by our failure to appreciate the pervasiveness of the material-semiotic coupling. Our own culture brings with it the ideological biases of a dominant class whose interests favor a view of the world as indefinitely exploitable materially and infinitely flexible culturally. We are an urban culture for whom agriculture, a primary site of the material-semiotic coupling, is distant and trivial. We are a machine culture accustomed to simple proportionality of cause and effect (not massive self-amplifications) and stable dynamics (not emergent self-organization). We are a culture reluctant to examine what we do culturally to and with organic bodies (our own, our children’s, our enemies, other species’; Lemke forthcoming-a). We are only beginning to realize that we are not the Lords of Creation, but the most expendable, vulnerable, dependent, recent extension of a far older, non-human planetary ecosystem (e.g. Lovelock 1989), and that our survival depends on enhancing, not exploiting, a system which takes no cognizance of our interests and values, except insofar as they long ago adapted to its realities. We are also only beginning to realize that we do not make history, and culture, exactly as we please, but only within the limits of a vaster, trans-human system, whom we cannot in principle observe or control.

What makes a system `complex’ in this sense? How are the dynamics of such complex systems as human communities and their ecosystem bases fundamentally divergent from intuitions about them based on the dynamics of simple, machine-like systems? Consider some examples of complex systems in this sense: a dust-devil (or a tornado), a cell, a developing embryo, a caterpillar-pupa-butterfly, a human organism, a living lake (or rainforest), a living city, an ecosocial system, the living Planet (Gaia).

The study of complex systems is now well advanced in physics and chemistry (e.g. Prigogine 1980, Prigogine & Stengers 1984, Jackson 1989, Harrison 1982) and is beginning to make progress in developmental and evolutionary biology, ecology, and geophysiology (see Weber, Depew, & Smith 1988; Odum 1983; Salthe 1985, 1989; Holling 1986; Lovelock 1989). What makes a system truly complex dynamically is not simply the number of variables (or `degrees of freedom’), but how these variables depend on one another, the pattern of their `couplings’. The more interdependent they are, both in numbers of interconnections and the strength of the interconnections, the less predictable the future of the system. When the couplings `loop back’ on themselves (e.g. changes in A produce changes in B, which produce changes in C, which in turn produce changes in A again), the system may grow in complexity, generating new global patterns and new information.

Physics first studied systems with many degrees of freedom but only weak coupling between its elements (e.g. gases with large numbers of weakly interacting molecules). For systems like this the first symptom of complexity already appears: the Newtonian symmetry of Past and Future is broken, dynamics proceeds irreversibly and uni-directionally into what we call the Future. Each separate internal interaction or collision of molecules is a simple system and could in principle be reversed in time, brought back exactly to its previous state with a finite amount of information and a finite amount of energy. But each collision produces correlations in the subsequent motions of the participating molecules, which now have further and further collisions, the correlations multipling rapidly toward a state which would require infinite information to be set into an exact reversal (Prigogine 1980).

This phenomenon of irreversibility was first formulated as the famous Second Law of Thermodynamics: that closed systems tend to the state of equilibrium, the state with the most probable values for any overall macroscopic property of the total system, corresponding to any one of a set of the largest possible number of thermodynamically equivalent distributions of the molecules. Any other state would be much less probable because there would be many fewer molecular combinations corresponding to it; random collisions would rapidly favor the equilibrium state. Equilibrium is also the most homogeneous, most symmetric, least diverse, coolest, lowest energy state. It is the final death, the endpoint of decay and decomposition: neutral, inert, exhausted, stable. Spontaneous thermodynamic change moves from the unusual, the specialized, the differentiated, the energetic to the generic, the uniform, the quiescent. From what is uncommon and improbable in world of random influences that destroy order and organization to the most probable state of no order, no organization; from states high in order and organization (`high negentropy’) to those high in disorder and disorganization (`high entropy’).

But in the real world many complex systems, and all the ones on our list of examples, do not behave in this way. A mass of air with a vortex (dust-devil or tornado) is more organized than the turbulent mass of air before the vortex formed, not less. A developing embryo goes from a state of lesser to a state of greater differentiation, away from homogeneity. Mature ecosystems are more complex and differentiated than immature ones, not less. The living Planet as a whole is today further from the state of equilibrium than it was 4 billion years ago, not closer to it. It would be very easy to predict the future of a culture, of an ecosocial system, if it behaved thermodynamically: it would disintegrate, collapse, become homogeneous and incapable of further change. Distinctions would be lost, diversity would disappear, decay would outstrip construction, useless wastes would be more common than useful resources. But in fact our history has veered far away from this path to the ecosocial death of equilibrium, placing many buffers between us and the long slide to ruin. How?

All our examples are open, not closed systems; they all exchange at least energy and information, and usually matter as well, with their exterior environments. The living Planet lives because energy flows to it from the Sun, is transformed by life, and returned to space as radiated heat at a lower temperature than it would be by a barren planet in the same orbit. The developing embryo (and child) feeds on the nutrients and organizational information of its external (mother) and internalized (DNA) environments, producing great amounts of waste heat and waste chemicals which must be safely conducted away. The city claims resources of energy and raw materials from its environment and exports back to that environment heat and solid wastes in quantity. In all these systems, the transforming processes (metabolism, chemical ontogenesis, urban production and consumption) are irreversible ones and generate entropy (disorder, matter and energy closer to equilibrium than they began), but the high entropy elements are excreted from the system into the environment, allowing a net increase in the order and organization of the system itself at the expense of its environment.

The unavoidable, irreversible thermodynamic processes that generate entropy (in the form of heat and waste) are called dissipative processes, and a system that keeps itself going (and perhaps becoming even more organized and differentiated) by the trick of importing energy, information, and resources from and exporting disorder to its environment is called a dissipative structure (Prigogine 1980, Prigogine & Stengers

1984) or dynamic open system (Lemke 1984). Dust-devils, cells, developing embryos, organisms, ecosystems, cities, ecosocial communities, and the living Planet are all dissipative structures. So also are fires, hurricanes, convection cells in heated fluids (e.g. Rayleigh-Benard cells, atmospheric circulation and oceanic currents), certain chemical cells that maintain concentration gradients and produce elaborately beautiful patterns (see Prigogine 1980; Berge, Pomeau, & Vidal 1984).

The flows of energy, matter, and information that maintain these systems in existence are thermodynamic constraints, they keep the system away from the path to equilibrium by supplying order and safely conducting disorder away. How can such systems come into being in the first place? Once in existence, how can they become even more organized and complex, actually moving further from the path to equilibrium? The answer again is strong coupling. This is most easily seen in the case of chemical reactions involving several different chemicals which tend to form loops of chemical reactions, with some of the initial substances eventually being reproduced in the course of subsequnt reactions, thus leading to even more possibilities for the chain of reactions to continue instead of eventually coming to a halt (the path to equilibrium). Such systems of coupled, looped reactions are called autocatalytic systems, and they lead to conditions in which the amounts of various substances and the rates of reactions using and producing them depend on one another in more than proportional ways (i.e. an increase of 10% in some amount or rate might lead to more than a 10% increase in something else, even, ultimately, in itself!). Mathematically, these are non-linear systems, and they do not behave according to the intuitions commonly found in a machine culture.

Non-linear, autocatalytic systems are complex in the second degree. They not only show irreversibility, they exhibit the phenomena known as bifurcation (e.g. Prigogine 1980, Prigogine & Stengers 1984; cf. `catastrophe’ in Thom 1975) and chaos (Gleick 1987, Jackson 1989). Essentially these are systems that can shift unpredictably from one metastable non-equilibrium state to another. When they are embedded in buffering, regulating environments (supersystems) they can shift to new states further from equilibrium by dynamical symmetry-breaking.

Irreversibility was already a breaking of the symmetry of time in fundamental interactions (which are time-reversible, not distinguishing a `past’ from a `future’). Bifurcating systems create for themselves possible states with less symmetry (in time, e.g. periodicity, or irregular rhythms, or in space, e.g. gradients or spatial patterns) than they began with. There are always several such states, and if the system spent equal amounts of time in each, the net result would be to restore the original degree of symmetry. But this is not what happens. As the possibility of the new states is reached (because of a build-up of some internal or external factor), random fluctuations determine that the system will become stable in one of these less-symmetric, less homogeneous, more specialized, differentiated, orderly, organized, further-from-equilibrium states. If the experiment is repeated, it might be another of these asymmetric states that is entered and which then persists, but in each case the system moves further from equilibrium.

Such systems are often called self-organizing systems, although it is important to remember that the organizing of the system is the result of interactions with the environment, not an internal and autonomous process. In each new state of the system

the internal dynamics are different: there are different rates of reactions, different amounts of chemicals built up, new effects on the environments and new environmental reponses, and new possibilities for still newer reaction pathways to come into being. In this way a new state of the system prepares the way for yet another bifurcation, yet another jump to a still newer state, even more organized and differentiated, breaking more symmetries of the previous state. Again, accidental factors may play an important role. New couplings of reactions may occur in this individual system and not in that, random fluctuations (internal or external) may influence the possibilities of subsequent jumps to new states. We now arrive on the threshold of a further order of complexity in dynamics, one particularly characteristic of organic and ecosocial systems: the order of epigenesis, evolution, and emergence.

 

Epigenesis, Evolution, and Emergence

Electrons and atoms do not age. They have no history, no individuality, no youth, maturity, or old age. An atom is already a compound, though not in our sense a complex, system. It has different states, but they are always the same set of possibilities. It does not know irreversibility, it is not a dissipative structure. It is stable. If you shift it to an alternative unstable state, it quickly returns to its original configuration and all memory of the excursus is lost. You cannot tell one atom of oxygen in its stable state from any other; the definition of the state itself specifies all variables, there are no supernumerary degrees of freedom left to record a past history and allow us to distinguish different individuals in the same state.

Complex systems are very different, they have so many more degrees of freedom than an atom, that there are always degrees `left over’ as it were to record history, even if it is only the correlations that reflect the history of past random collisions. Complex systems are individuals and they have a history; it is possible to construct a continuity of individuality from before to after an interaction that changes the system in some way. But if such systems have undergone a series of bifurcation jumps to new lesser-symmetry states farther from equilibrium, then it is not possible to predict (or model in any way) these future states from a knowledge of prior states, except by recapitulating the intermediate states, i.e. the entire developmental sequence of bifurcation jumps, leading to that future state. Self-organizing systems thus have a second, invisible history: not just marks of wear and tear, the accumulated memory of past encounters, but the developmental trajectory of changes in their dynamics by which they came to the more organized, less symmetric state they are in.

In a profound sense, complex systems which develop in this way, including both human organisms and ecosocial systems, are temporally-extended entities. The system, as an individual entity, cannot be defined at one moment in time, because the dynamics which maintains it in being must occur over time. In each instant, it is dead; only over time is it alive. So much is true for any dissipative structure, but a truly developing system cannot be defined even over an interval of time limited to one stage in its developmental career, because its constitutive dynamics will be quite different in later stages. Only the system extended in time along its complete developmental trajectory, from formation to disintegration, from conception to decomposition, is a properly defined theoretical entity. We will refer hereafter to the developmental trajectory entity, meaning the

system-over-its-lifetime, when necessary, to emphasis this new perspective. The caterpillar-pupa-butterfly is one individual developing system, as is the embryo-child-adult-dotard. The notion of trajectory-entities in this sense allows us to formulate new, genuinely dynamical definitions of ecosocial systems, cultural formations, language dialects, corporate institutions, and even social individuals.

With the notion of developing systems we reach a new threshold. But before we go beyond self-organizing dissipative structures, it is worth noting that all such systems exhibit a common thermodynamic outline to their trajectories of development (see Salthe 1989). Whether we are speaking of hurricanes, embryos, organisms, or ecosystems, there is a common sequence of developmental stages. In the ascendant stage (or phase) the system is dissipating energy, producing heat and wastes (entropy), at a maximal rate in proportion to its total mass, and its internal organization and order are increasing at the maximum rate. This rate of generation of disorder (exported to the environment) and order (accumulated in the system itself) gradually slows as the system passes through its various developmental bifurcations, moving further from equilibrium, until some limit is reached, and a metastable state develops (mature phase) with minimum entropy production consistent with maintaining the mature organization of the system. Finally, there may be a senescent phase in which an overly self-regulated dynamics becomes vulnerable to external disturbances, eventually degrading and finally decaying back toward the path to equilibrium and death. The most complex developing systems may be able to avoid terminal senescence, as we will see.

The existence of such a generic developmental trajectory points the way to a new strategy for modeling complex systems. For any given system, it is not possible to anticipate bifurcations and predict dynamical futures beyond the current stage (or even whether there will be a new stage), but if the system is of a recognizable type, then there is a good chance that it will follow, at least up to a point, the typical developmental trajectory of its kind. Type-specific developmental change is predictable from a knowledge of the type. This is the basis of embryology, and of the prediction that most caterpillars, if they survive under more or less normal conditions, will eventually be butterflies. But how does it happen that developmental trajectories as specific to a set of ecological conditions as those leading to butterflies can become fixed and repeatable? This is the next order of complexity in dynamics, that of epigenesis.

An epigenetic system is a developing system that recapitulates the major stages of a type-specific developmental trajectory. It is a system that develops according to its kind, recapitulating a sequence of bifurcations in the dynamics of its type that may have evolved over many generations of its predecessors. I hope it is clear that while we have for some time now been using the language of living systems, that at no point in the specification hierarchy (Salthe 1989) we have been defining (each type of system a special case of the previous: complex systems with irreversibility, dissipative structures, developing systems, epigenetic systems) is there a clear transition to Life, as such. Hurricanes are alive in many significant ways; so is the Planet as a whole. Organismic life as we know it is based on a very specific strategy (DNA-mediated epigenesis), but ecosystems are also alive and use a different strategy. What is special about the class of epigenetic systems is that the developmental trajectories of individuals recapitulate a prior evolution of the trajectory of their type.

The terms `development’ and `evolution’ are used loosely and often interchangeably outside (and even inside) biology. Because they are paradigms of different modes of change in the study of complex systems, however, it is important to separate them. Individuals develop; types evolve. Individuals also individuate; that is, the developmental trajectory of an individual system recapitulates that of its type only in general: in many specific ways it is unique, reflecting its own individual history. In particular, an individual system may deviate from the type-trajectory in a way which can be passed on and recapitulated by future developing systems: it may contribute to the evolution of the type. What evolves is the developmental trajectory of the type (and not, actually, the type as such). Evolution occurs when individuation leads to a new dynamical stage (through a new bifurcation) which can be recapitulated, and when the new trajectory actually is recapitulated in the developmental trajectories of a signficant number of successor systems.

How is recapitulation possible? Epigenesis adds one element to development itself: a guiding environment shared by different individual systems and relatively slowly changing compared to the lifetime of these systems. The sequence of bifurcations, of development, cannot be left entirely to chance, to random fluctuations, if there is to be recapitulation. Random fluctuations must be harnessed and guided by an external source of information, regulation, control, and that can only reside in the environment of the developing system. An adequate analysis of a developing system must not only be extended in time, but it must examine system-environment interactions: it must extend to the immediate supersystem that contains both the system under focus and its immediate environment (cf. Lemke 1984). Dissipative structures and their environments are necessarily mutually interacting. The sequence of bifurcations will depend strongly on environmental conditions. If many individual systems develop under the same environmental conditions, the odds are that they will undergo similar sequences of bifurcations. Developing systems can and do also modify their environments (often for the worse, by exporting disorder into them), but they are dependent on these environments for energy, material resources, and information-regulation-constraint.

The next step is simple: a new bifurcation in an individual leads to an effect on the environment that favors similar bifurcations in other individuals: epigenesis is born. A series of `accidental’ dust-devils in a narrow defile might erode landscape surfaces in a way that produces contours which favor the formation of locale-specific dust-devils. Globules of organic polymers engaged in autocatalytic chemical reactions might modify the surrounding silicate clays in ways which tend to favor their latest chemical innovations in successor globules. In each case, along with epigenesis comes a supersystem and a hierarchical relation of system and supersystem. That hierarchical relation is one of scale (cf. Salthe 1985, 1989 on scalar hierarchies), in which the supersystem is more stable, changes more slowly, and exerts a regulatory influence on the dynamics of the now `sub’ -system. In the case of organismic lifeforms, the relatively stable `environmental’ molecules (RNA, DNA) were eventually internalized, incorporated into the supersystem which became the modern cell.

But epigenesis depends only on the existence of the subsystem-supersystem regulatory scale-hierarchy relation and the possibility of innovations in subsystems being recapitulated through their long-term `memory’ by the supersystem environment (and not specifically on the DNA strategy). Epigenesis is simply development under an environmental guidance that enables the recapitulation of type-trajectories in individual

development. Moreover, the hierarchy principle is automatic: because dissipative structures are already a product of system-environment interactions, a supersystem is always already implied. This account of things (like many of the accounts of the origin of organismic life) is a bit backwards: there have always been supersystems, there have always been ecosystems, there has always been a planetary dynamical system. Particular self-organizing units always came into being in the context of such supersystem environments. Life did not begin with micro-organisms that eventually got together to form ecosystems that eventually united into Gaia. There was always Gaia, even pre-biotically, and there were always the precursors of ecosystems. What has happened in the history of the planet is that new intermediate levels of organization have emerged between the Gaia-system and her autocatalytic and cross-catalytic molecular subsystems. Ecosocial systems and the human cultures they sustain form one of those intermediate levels.

All epigenetic systems belong to regulatory subsystem-supersystem hierarchies from the molecular to the planetary. But organismic lifeforms are not the only epigenetic systems, there are also ecosystems, and they take us on to the final principles of complex dynamics needed for understanding cultural and social change.

 

Ecosystem Dynamics

Living forests, lakes, and cities are also epigenetic systems, but they have evolved somewhat different strategies from those of organisms. Organisms and ecosystems are both larger-scale supersystems constituted by and acting to integrate and regulate the smaller-scale subsystems they contain. Organisms integrate the processes of organ systems, organs, tissues, cells, and intra- and extra-cellular body chemistry down to the molecular level. Ecosystems integrate the processes of interaction of organisms with each other (within and between species) and with the flows of matter, energy, and information through the total system, including solar radiation and heat flows, water and nutrient flows, and hydrologic, atmospheric, and geologic, processes on local and larger scales.

Organisms show `planned obsolescence’, they enter a developmental phase of terminal senescence and die. Most higher organisms seem to have a fairly definite maximum lifespan, after which they quickly return to the path to equilibrium. This is associated with their strategy of individual reproduction: organisms are like autocatalytic reactions, they multiply themselves to the limit of available resources or until they are regulated by the supersystem (e.g. by predator population increases). If such `breeder’ lifeforms did not die, they would soon preempt all available resources for their progeny and the further evolution of their type would be inhibited. At the ecosystem level of organization, however, things are rather different (Odum 1983, Holling 1986, Schneider 1988).

Ecosystems do not seem to die of old age; neither do they directly reproduce new individual ecosystems. The relation between individuation, evolution, and recapitulative development is different at this level, but it exists nonetheless. Ecosystems do show a form of type-specific recapitulative development, known as ecological succession. A newly opened area (a new volcanic island, a burned-out forest, abandoned farm-land) is

first colonized by one group of species that form an ascendant phase ecosystem with its own stage-specific dynamics. This ecosystem tends to spread rapidly, with fast-reproducing, short-lived species; it also alters the soil and local environment generally in a way that is favorable to its replacement (succession) by another group of species that forms a later phase of the ecosystem’s over-time developmental trajectory, and which can flourish in the conditions created by the first phase. This continues with a slowing in the rate of growth and the rate of dissipation and accumulation of structure; there are more complex couplings of species and nonbiotic elements, longer chains for the cycling of nutrients, more stored resources, etc., forming a mature ecosystem.

But instead of heading on into terminal senescence, two things will have happened instead. The system may enter a stage of post-maturity in which it is not as resilient as previously and is more vulnerable to external disturbances, but at the same time it will have grown `patchy’ with a mosaic structure on many scales of small regions in which the dynamics are distictively different. When a great oak or redwood finally dies, when a small fire burns out a part of the forest, when a storm damages part of a coral reef, when pollution degrades the environment in part of a lake, a mini-succession will begin again in that patch, progressing faster or slower depending on proximity to other mature patches, which species’ propagules get there first, and what the local soil, light and water conditions are. The natural topographical variations in soils, and for marine ecosystems the natural patchiness of nutrient flows and plankton populations, also insures that ecosystems are everywhere `patchy’, mosaic aggregates. Stresses on patches may even cause a retrogression in the successional developmental sequence.

The result is that ecosystems are mixed-age aggregates. They consist of parts at different ages or stages of successional development. And they consist more generally of little mini-ecosystems with slightly different mixes of species, or even different species in the same functional niches. Ecosystems do not, like organisms, reproduce new individuals with a distribution of variation in characteristics; they contain this diversity within themselves in simultaneous mosaic patches. Not only age, or successional stage, but every other characteristic of an ecosystem is present within it with a distribution of various values at various scales. Ecosystems are mixed-age, mixed-character mosaic aggregates.

Every ecosystem is an individual, and in a looser sense than for organisms, a member of a type. Its successional trajectory is not as rigid as that of an organism, but it has a recognizable outline. The succession of ecosystems is not under as tight a regulatory control as is the development of an organism, because the ecosystem is not itself part of as highly organized a supersystem as is a developing embryo. Ecosystem types are not as well-defined as organism species are; they are `fuzzier’ types. Ecosystem types certainly evolve (in our strict definition of evolution, it is their actually their successional trajectories which evolve), and may do so very rapidly. The same species may form a different ecosystem if the dynamics of interaction in that system are different, and innovation (new couplings, interactions, effective bifurcations in the population or energy and matter dynamics of the patch-subsystem) in a relatively young patch as it undergoes its partially unique succession can then spread in time through the territory of the total ecosystem, or at least be recapitulated in other patches, with evolution taking the statistical course of the most frequent pattern in future patches.

Ecosystems do not reproduce, but they do spread (and survive, and `dig in’, becoming more resilient and stable occupants of their territory). The criteria of success, of `fitness’ for ecosystems cannot be posed in the same terms as for breeder lifeforms (whose reproductive fitness is of course itself very critically a function of their ecosystem contexts). For ecosystems the criteria of success include: persistence (metastability, resilience), adaptability, creation of a successful relation with an environmental supersystem (for resource inflow, waste outflow, buffering against disturbances), colonization potential, and optimization of mosaic age- and diversity distributions.

Ecosystems follow an epigenetic strategy that fosters the recapitulation of type- (and microtype-, i.e. `patch’) -specific successional trajectories by both internal memory (the total ecosystem being a slowly-changing, regulatory, environmental supersystem with respect to its patch subsystems) and external memory, such as modification of soils, water tables, local landforms, microclimate, etc. In this, in their lability to rapid local evolution, in their mosaic diversity, non-reproduction, and criteria for success, they seem to present a much more appropriate model for the dynamics of cities or human social systems than do organismic systems. This is because human social communities are, or more accurately, are part of ecosystems.

 

Ecosocial Dynamics and Semogenesis

We can now begin to fully interconnect our three basic arguments: that human sociocultural systems are essentially systems of social practices linked into the historically and culturally specific semiotic formations in which they take their meanings; that these practices are simultaneously material processes in a complex, hierarchically organized, developing and evolving ecosocial system; and that the dialectical relations between the semiotically- and materially- based couplings of these practices/processes are the basis for a general ecosocial dynamics.

Cultural dynamics is one aspect of the total dynamical complexity of what we are calling ecosocial systems. An ecosocial system is a human social community taken together with the material ecosystem that ennables, supports, and constrains it. An ecosocial system is an ecosystem, with all the characteristics and properties of ecosystems, but it is a more specified type of ecosystem: one that includes a community of organisms of our species and in which therefore the material interactions of its elements (people, other species, resources, material and energetic processes and flows) are biased, constrained, and organized, in part, in accordance with social semiotic formations. These formations are constellations of actions-as-material-processes organized in terms of their meaning relations as social practices in a human community.

The total ecosocial system includes not only human organisms and their interactions with one another, but all the material elements which act on, in, and through humans and which humans act on, in, and through. It includes all the other species with which we are co-dependent and with which we have co-evolved, including our food species, our diseases and parasites, our symbionts and co-dependent micro-organisms, and their webs of interdependent organisms and material and energy flows. An ecosocial system includes buildings and tools, cultivated fields and soil bacteria, generating stations and bread molds. It includes landforms and marine nutrient flows, atmospheric circulation

and solar radiation levels. It includes manufacturing and waste production, education and intercourse, politics and warfare. And it is a single, unitary system in which the dynamics of processes of human social interaction are not in principle or in practice separable from the dynamics of the rest of the ecosystem, except that cultural practices represent a second level of organization of material processes according to relations of social meaning. An ecosocial system is simultaneously a material and a social-semiotic system.

Ecosocial systems show ecosystem organization in both their cultural-semiotic and their material-ecophysical dynamics. They are foremost, in both respects, not systems of things (organisms, nutrients; subjects, signs) but systems of processes (gene exchange, predation; communication, production). They are systems precisely insofar as these processes are coupled: linked, interconnected, interdependent. They are complex, open, dynamical, dissipative, self-organizing, developing, individuating, epigenetic systems, organized in a hierarchy of levels in which subsystem development and individuation is regulated by supersystem dynamical maintenance, and in which supersystem resilience and adaptability is insured by subsystem variety and lability to new patterns of cross-coupling. Their hierarchical structure arises from the interpolation of new intermediate levels of dynamical organization as new patterns of process-coupling (directly or through the coupling of social practices in cultural formations) lead to symmetry-breaking and new dynamical states (emergent structuration). Ecosocial systems are mosaic aggregates of subsystems (`patches’) of differing developmental age, composition, and coupling patterns.

Within this general model of ecosocial dynamical systems, let us consider in more detail the dynamics of social practices and formations, of cultural systems of meaning. Of all these, Language has traditionally been regarded as the least materially coupled and has been most often recruited as the paradigm for an autonomous semiotic dynamics. But Language-as-system is an abstraction from language-in-use, from the social practices that employ and deploy the resources of that system. If speech seems materially a matter of a breath and negligible energy, and writing of only the infinitesimal energy and entropy of inkstains on paper or magnetic domains in an electronic memory, the same could be said of the DNA genome that guides and channels the much larger energies of the chemistry of embyrogenesis. The genome, like Language, has evolved to be what it is (and continues to evolve) exactly by serving this function, and thus precisely in and through its couplings to those larger processes. It is just the same with Language, through which we construe the meaning relations that tie together the social practices of every semiotic formation, so that we may learn to make sense of and with them. In what we do with Language (and with every other semiotic resource system), strong couplings are made between the material processes through which we enact all other social practices, many of which engage and entrain substantial flows of matter and energy. Discourse formations construed in Language guide the social practices of our architecture and our engineering, our agriculture and our industry, our choices of foods and mates, allies and enemies.

The cultural dynamics of Language cannot be independent of the uses to which language is put, but rather arises directly out of those uses. The cultural dynamics of Language-as-system, the processes of Language change, represent the effects on the overall semantic resources of a language of the new uses to which it is put, register by register, function by function, situation by situation. This putting-to-use is always a

putting to use in the material doings of an ecosocial system. If a language has an inertia, it is the inertia of its use by many organisms, a dynamical inertia grounded in neurological processes and the material social practices of producing and interpreting stable, visible material images. If Language has a momentum, it is the dynamic momentum of changes in the social practices of language use.

The notion of a `language’ is not only an abstraction from use, it is an abstraction from the empirical diversity of language in use. A language is a mosaic aggregate of its dialects and sociolects, on all scales of a hierarchy of loosely integrated subsystem `patches’ (cf. `speech communities’) from those of widely used dialects down to individual idiolects. This principle of mosaic diversity applies equally to registers and to discourse formations, and with respect to the latter, ecosocial heteroglossia encompasses the diversity of language use across different age-groups (cf. mixed-age mosaic ecosystems), genders, social classes, political persuasions, etc. And in every `patch’ of language use, in every functional `niche’ (situation-type) of language use, the pattern of use is changing: developing, individuating.

We need look no further than the phenomena of creolization to observe recapitulative and individuating development of a language `patch’. Just as extreme stress on an ecosystem, measured by decoupling of processes and flows, by loss of differentiation and return toward greater symmetry and homogeneity, leads to a reversal of the stage sequence of succession (Schneider 1988), so the restriction of the use of a language to only a few situations and to speakers who can only learn the language in those situations (`pidginization’) leads to a loss of functional and semantic differentiation (and so to phonological homogenization as well), a simplification of language not unlike the earlier stages of language-learning by children or non-native speakers. A `language patch’ is cleared, or at least pushed back to a more primitive state of development. But it is a patch in an ecosocial system; the diverse activities of the community are still being enacted, there is still a rich field of differentiated functions and contexts into which the `pidgin’ can spread. As it does so, its symmetries are broken, unitary forms become differentiated and multiply in meanings and uses, a `full-service’ language rapidly (in one generation) re-emerges. It is not identical to the original language, for it does not recapitulate a precise epigenetic trajectory like an organism (cf. a child learning the mother-tongue), it is an individuated patch, perhaps a new dialect of a larger language family (depending on the degree of integration of the local community within the larger ecosocial system of that family).

The stages of development of a creole dialect, while they may not recapitulate the history of the original language in detail (except in the presence of DNA-like environmental `templates’ to bias the development) will still proceed by a series of symmetry-breaking bifurcations, each the necessary predecessor of the next. Each later state of semantic and functional differentiation in language use patterns must be prepared for by prior developments that enable the meaning contexts to be created (and recognized) within which the dynamics can give rise to the subsequent ones. In this process, greatly accelerated in speed (cf. rates of succession or ontogenesis vs. those of evolution and phylogenesis), we see the same interdependence of social action patterns and semantic resource development that occurs in the normal course of language change.

The processes of symmetry-breaking common to all developing systems with complex non-linear, autocatalytic dynamics occur in developing ecosocial systems (as wholes and more obviously in each patch) and are reflected in the successive symmetry-breakings of normal linguistic change. Some of these have been documented for modern English by Halliday (in press-a), who has identified the pattern he calls semogenesis, in which a progressive semantic differentiation in the language system is built up by the context-dependent splitting of previously unitary semantic features and their subsequent independent recombination. He also indicates how this same process, writ large, leads to the emergence of a true grammar as an intermediate level of semiotic organization between functional situations and utterances in the ontogenesis of speech in children. It is very likely that this is also the historical sequence of the origins of Language as we know it.

The total semodynamic cycle is far richer and more complex than the core semogenesis process itself. It is a cycle in which differences create (or enable the creation of) further differences, in which the frequency of occurrence of a feature proceeds from being equal across contexts to being unequal (symmetry-breaking), and in which the availability of new semantic features makes possible the differentiation of new contexts in which in turn further semogenesis can occur. An essential part of this cycle is the existence of pre-semiotic features of events or situations: material differentiations which do not yet have cultural significance, but which can enter the semiotic system as new features (cf. Lemke 1984, forthcoming-b).

Bifurcations in the material ecodynamics of an ecosocial system generally lead to a breaking of the symmetries that existed in prior states of the system, so that new differences are potentially distinguishable: what was formerly one single context may now be separable into two somewhat different contexts. This difference of contexts may now be used to `pry apart’ a formerly unitary social practice, if distinguishable (but previously type-equivalent) varieties of the enactment of the practice begin to co-pattern regularly with the difference in contexts.

Every material instance of a form or type (be it a context-type, a practice-type, etc.) exhibits both the criterial features of the type and additional incidental features, which do not matter as regards the taype. If some of these incidental features begin to co-pattern with different context-types, not just in isolated instances, but regularly (owing either to material connections between them or to semiotically constructed ones) and recapitulably, then the former type, previously symmetrical as between these contexts, is now split into subtypes by the formerly incidental but now criterial (for the subtypes) features. Material symmetry breakings and couplings can lead to semiotic ones, and vice versa. Differentiations of contexts can lead to differentiations of practice types, and vice versa. When features (of practices and/or contexts) uniformly co-occur (perfect redundancy) across all wider contexts, they are not semiotically separable as distinct features, but when they begin no longer to do so in some contexts, a semogenic process may begin in which they become separable in all contexts in which they occur. As their degree of redundancy (probability of co-occurence) falls from maximal toward zero, they become independent resources of the meaning system (cf. Halliday in press-a, in press-c; Nesbitt & Plum 1988), increasing its information carrying capacity.

But at the same time there is an opposing tendency, since in at least some contexts there must be greater redundancy for there to be cultural formations: when all combinations

are equally probable in all contexts, there is no culture. Redundancies and finite probabilities of co-occurence define formations and increase the total organizational information of the ecosocial system. There are furthermore `semolytic’ processes by which previously distinguished features fuse, fall into disuse, or do not continue their distinctive associations with contexts (Lemke, forthcoming-b). There is thus a `semodynamic cycle’ in which new distinctions of meaning, new resources, and new formations are continuously created and destroyed, all as part of the total material-semiotic dynamics of the ecosocial system.

The net result is that the meaning potential of the system of semantic (more generally, semiotic) resources increases (as new features are added, and become combinatorially independent, so that all possible combinations of features may occur), while at the same time the total organization of the semiotic behavior of the community increases as the probabilities of combinations become increasingly context-dependent (cf. the `meta-redundancy relations’ described above).

It is perhaps easier to see the developmental dynamics at the level of semiotic formations rather than at that of the necessarily more abstract semiotic resource systems. Consider some system of technological practices, that is, some interlocking cultural formations of technology-using social practices in agriculture, manufacturing, warfare, etc. Now suppose that a nuclear holocaust had destroyed the material base of the technologies, or caused a retrogression to a more primitive level of technology, but that the `template’ (say a discourse formation, the `knowledge’ of the technology) still existed. This is like the classic problem of the modern man in a stone-age society, or the traveler-back-in-time seeking to make use of advanced technological knowledge in the absence of the material base. Some short-cuts may be possible, but by-and-large the `succession’ in this `patch’ will have to recapitulate the historical (or at least the obligatory developmental) sequence of technological developments. Each development makes possible the next. In historical dynamics, each development makes possible new social practices, which combine into new cultural activities and institutions, with new needs and interests, which spur new technologies to meet those needs, built on existing technologies designed to meet earlier needs, and so on.

Ecosocial systems are hierarchically structured across many scales of organization. This does not mean that they are in any sense `authority hierarchies’ or that authoritarian social organizations are more natural than democratic ones. It means that each level of organization going `up’ the hierarchy is in fact a larger-scale, more slowly-changing supersystem coupling and integrating smaller-scale, more rapidly-changing susbsystems at the next level `down’. A nation is a supersystem of its cities and provinces; the global economy is a supersystem of national economies. In authoritarian social organizations, elites become smaller as one goes `up’ the hierarchy; in ecosocial systems, the higher levels are orders of magnitude larger.

Ecosocial systems, and each patch and subsystem within an ecosocial system, is an irreducibly temporally-extended entity: it is a system of coupled, interdependent processes, including exchanges of matter, energy, and information (entropy) with its environment, which are constitutive of its existence and whose nature changes across the system’s developmental trajectory partly as a result of the effects on the system and the environment of the processes at a prior stage. Ecosocial systems generate their own futures (at each level in interaction with an environment that may be partly a regulatory

super-system): they create conditions which lead them to change, and they create at each developmental stage possibilities for the next stage which are not in principle predictable. Accidental, unique, historical configurations will often determine which of several possible `branches’ the developmental trajectory of the system will take. Larger-scale, longer-term conditions will determine which of the recapitulable innovations of an individuating system will in fact contribute to the evolution of the type.

At any given time the characteristic cultural patterns of action of a community must be enacted by material processes, by actual human organisms in interaction with other ecosystem elements. Each enactment of a ritual, each performance of a song, each making of a tool, each writing of a sonnet will be unique and different, but it will also re-enact criterial features common to the type, to the cultural formation, the social practice. Other, initially incidental features, may in the course of cultural change become newly criterial ones for an evolved type. There is an essential dialectic between types and their `tokens’, between abstract practices and formations and their individual instances. This is a dynamic dialectic: it leads to change, it mediates the process of change. Types (semiotic elements) change because they must be instantiated as tokens (material elements), and tokens of complex systems are always unique.

Consider this dialectic not in relation to an `object’ but to a process, or more typically for a cultural system, an `event’, an `activity’. The abstract type here is the actional semiotic formation, the `action genre’, `activity type’, or `participation structure’. It calls for actors in various roles to perform various actions which couple the material processes of organisms and other ecosystem elements together in particular ways. It defines the roles and action-types which constitute it; it construes them, in terms of criterial features of the material processes, neglecting incidental features, as socially meaningful, and it is in terms of meaning-relations that it prescribes the couplings of actions/processes. But each enactment of the event will be different. The organisms performing its key roles are themselves developing systems, with unique histories of participation in other actional formations of the culture; each is a unique constellation, with a slightly different view of the current role and how to perform it. Participants are changed by their participation in cultural events, and the cultural events as types may change as a result of their enactment by different unique participants. The developmental process, the semodynamic process, is at work here. Each event, each participation, creates conditions which may lead to further change, in individual history (of organisms, of communities) and in type evolution. Ecosocial system are not stable; they create the conditions for their own change.

 

Prediction, Control, and Responsibility

Type-specific, recapitulative development is both lawful and predictable. Evolutionary change is lawful, but not predictable in detail beyond the short-term in which environmental constraints set conditions for the spread or extinction of new coupling patterns. Individuation is neither lawful nor predictable; it is the source of new variety in the history of the system, unique and accidental.

Recapitulative development is only as faithful to the type trajectory as accidental conditions and the epigenetic regulation of the material `template’ allow. Actual system

trajectories represent the combination of recapitulative development with individuation. The `template’ in ecosocial systems is not a stabilized internal DNA, but rather the persistence in the system’s environment of the patterns of coupled processes that biased the developmental changes of others of its type, perhaps reinforced (or modified) by these predecessors’ cumulative impact on the environment. When an individual, or a community, `learns’ a new pattern of behavior typical of their `kind’ (e.g. how to program a computer), there will be a certain recapitulation of the historical stages of skill differentiation that led to this activity-type, guided and in part provoked by cultural discourses and other formations. But each individual and community will nonetheless develop its own unique approach, and some of these approaches will spread and contribute to the evolution of this social practice, while others will die out. At any given time there will be a mosaic of system `patches’, each with a variant approach, and the variants will most often arise in `juvenile’ patches, where the most signficant differences will arise from deviations from the previous type-trajectory early along its path (cf. neoteny in evolution, e.g. Gould 1977, Montagu 1981).

Evolutionary change is lawful in that at any given time it is possible to specify the conditions that favor or disfavor the persistence or spread of a particular innovation. The evolution of the type is determined, strictly speaking, by the changes in the frequency distribution of recapitulations of the various variants of the type. But this short-term predictability of evolutionary change depends on the fact that the relevant environmental conditions are relatively slowly- changing, which is usually insured by the hierarchical structure of the supersystem within which subsystem evolution occurs. Long-term evolutionary change is not in principle predictable because of the development of the supersystem (i.e. its individuation). The dynamics of the supersystem are self-altering: it creates the possibility for its own change, for new couplings of its subsystems, for new couplings to exterior systems, for externally driven bifurcations to new states. Moreover, the changes at lower levels add new coupling-scheme possibilities, and where these are autocatalytic, major and rapid change in the supersystem can take place which is not predictable because the precise nature of the accidental changes in subsystems which the new dynamical scheme exploits cannot be known in advance.

There is, however, one clear trend in what we may call the (non-recapitulative) development of an entire system hierarchy: progressive hierarchical structuration, through the emergence of new intermediate levels of organization. This accomplishes a tighter integration of the higher levels of the total system (through couplings among these intermediate-level subsystems). The emergence of a tightly interdependent global economy, and of regional, multi-national economic federations are instances of this trend. At the same time, the ecological interdependence of all the subsystems of the living Planet is growing greater as well, and these two trends will eventually combine to produce a bifurcation-shift in ecosocial development: either a catastrophic retrogression in both, or a significant re-organization of both. In fact, it is often observed in ecosystem dynamics that a retrogression must preceed a re-organization in order for the system to `back up’ to a branch-point from which the new organizational dynamics is accessible. On the global scale, the less catastrophic alternative strategy of a new developmental trajectory being pioneered in an isolated patch is not available (unless we imagine it happening, as ultimately it must, in an extra-terrestrial colony).

Apart from the question of predictability, there is the issue of control. We can and do make history, but certainly not just as we please. The issue is more complex however from the perspective of ecosocial dynamics. If social and cultural systems were relatively autonomous, then we might imagine that the cultural future at least was mainly up to us, but what must be controlled, if controlled it can be, is the whole of an ecosocial system, not culture or social organization alone.

There can be no question of long-term control from the present; the emergent properties of a developing-evolving system preclude this absolutely. Only a continuing, adaptable, long-term effort on the same time-scale as the control sought could succeed, and clearly any such subsystem would quickly be entrained in the total ecosocial system, becoming a part of what it seeks to control. The same is true for short-term control by any individual or group, which is necessarily already a subsystem of the ecosocial supersystem. Can a subsystem successfully regulate the supersystem of which is it a part? We do not speak here of accidental influence; the individuation of the supersystem is of course vulnerable to internal events. But regulation is more than impact; it is a governing, a systematic capacity to shift the far larger processes and energies of the supersystem by critical manipulations on a much smaller scale.

Such control-from-below is not in general possible in self-regulating, hierarchically organized systems because of the great differences in scale between levels (cf. Salthe 1985), and because such systems do not evolve with sensitive vulnerabilities to subsystem processes (ecosocial selection favors robustness, resilience, and even meta-stability with respect to internal fluctuations). There are however certain special conditions under which a developing, self-organizing system becomes vulnerable to otherwise negligible influences. When the system is at a critical bifurcation point, when conditions are such that either of two (or occasionally more) dynamical configurations are newly possible for the system, its self-regulation is as it were suspended, and it becomes extremely sensitive to small fluctuations (Prigogine 1980, Prigogine & Stengers 1984). Under these conditions, small perturbations from much smaller scales in the hierarchy may become greatly amplified, and coherent global effects can result, including the determination of which branch the system’s further development takes.

In an extremely complex case such as an ecosocial system, it is possible that there are always numerous bifurcation possibilities `available’ at various intermediate levels of scale (fewer at higher levels). Coherent action by many subsystems, linked through communication, can affect supersystem behavior, especially near these critical branch-points, but also to a lesser degree away from them. Finally, the kind of action most likely to open up new dynamical pathways for the system is a reorganization of the coupling scheme: the linking of processes/practices not previously linked, or the de-coupling of those that formerly were. This is also true of the coupling of whole subsystems. Such actions, semiotically, correspond to re-definitions of equivalence classes and relations of alliance or opposition: to making semantic distinctions not previously made, to combining thematic elements not previously combined, and thus to making conceivable actions that link processes or subsystems not previously linked. In some cases, again, it may be necessary to de-couple before re-coupling in a new pattern, and it may be only in newer, younger, developing subsystems that the new dynamical patterns can first be effected.

The meaning systems of a culture, in the sense of the metaredundancy relations among social practices (describing which ones are linked in semiotic-cultural formations, in particular activity-types, and which are not), both enable meanings to be made, meaningful social activities to be enacted, and enjoin the making of certain meanings, certain connections. Where every combination is equally likely, where there is no differentiation according to situation or context, there is no meaning. And where there are meanings enabled, there are necessarily also meanings disabled (cf. Lemke 1984 on disjunctions). The panoply of meaning relations that define a culture is a figure against the ground of meaning non-relations, gaps that are not even seen as gaps. New coupling schemes of social practices (and so of material processes as well) that fill these gaps, that make meaning in the interstices of culture, in the dark places whose emptiness of meaning defines the boundaries (and so the potential growing edges) of what is meaningful, are especially likely to contribute to shifts in ecosocial organization at some level.

Mature ecosocial systems ward off terminal senility by the incorporation of a mixed-age mosaic of alternative-succession patches which serve as a reservoir of diversity. Some of these patches may even serve counter-regulatory functions, preventing the system’s self-regulation from becoming overly rigid (and so unable to adapt to environmental changes). If not large-scale patches, then subsystems at some scale can be expected to fill this function. In ecosystem dynamics it has been noted (Holling 1986) that some ecosystems never reach a quasi-stable `stationary’ state, because there are no stable values of the various population and other parameters which the system’s dynamics will continue to regenerate. It remains in a state of continual dynamic disequilibrium, with all its parameters fluctuating (sometimes in erratic cycles, sometimes chaotically), continually seeking a steady-state it can never achieve. In such a system (and ecosocial systems are surely like this, at least in parts) no one species or dynamic coupling scheme ever `wins’: all co-exist uneasily, in endless competition, with the result that the total diversity of the system remains higher than in any possible stable configuration.

In these terms, we also serve, who obstruct stability, who contravene tradition, who say and do the forbidden. We do not know enough yet to identify the critical moments when our small influence might be amplified and guide the course of systems far larger and more complex than ourselves, but we can offer alternatives, even if only in small patches, and we can study the conditions of their survival potential relative to others. We can also de-couple systems of practices and processes that reduce the survival potential of all alternatives. The great intellectual task of the next century, already begun, is to unmake cultures that deny the unity of ecosocial systems and remake cultures that celebrate it. For this task we will need guiding theories far better developed than what I have sketched here, but I believe that the theories we need will make use of many of these fundamental principles.

 

References

 

Bakhtin, M. (aka V.N. Voloshinov) 1929. Marxism and the philosophy of language. 1986 edition. Cambridge, MA: Harvard University Press.

--- 1935. Discourse in the novel. In M. Holquist, Ed. The Dialogic Imagination (1981). Austin TX: University of Texas Press.

--- 1953. Speech Genres and Other Late Essays. 1986 edition. Austin: University of Texas Press.

Berge, P., Pomeau, Y., & Vidal, C. 1984. Order within Chaos. New York: Wiley.

Bernstein, B. 1981. Codes, modalities, and the process of cultural reproduction. Language in Society 10 (3): 327-364.

Eco, Umberto. 1976. A Theory of Semiotics. Bloomington IN: Indiana University Press.

Fawcett, R. 1980. Cognitive Linguistics and Social Interaction. Heidelberg: Julius Gross Verlag.

G?, K. 1931. On Formally Undecidable Propositions. 1962 edition. New York: Basic Books.

Gould, S.J. 1977. Ontogeny and Phylogeny. Cambridge, MA: Harvard University Press.

Gregory, M. 1967. Aspects of varieties differentiation. Journal of Linguistics 3: - .

Halliday, M.A.K. 1976. In G. Kress, Ed., Halliday: System and Function in Language. London: Oxford University Press.

--- 1978. Language as Social Semiotic. London: Edward Arnold.

 --- 1985. An Introduction to Functional Grammar. London: Edward Arnold.

 --- 1988. On the language of physical science. In M. Ghadessy, Ed. Registers of Written English. London: Pinter Publishers.

 --- in press (a). How do you mean? In L. Ravelli and M. Davies, Eds., Selected Papers from the Seventeenth International Congress of Systemic Linguistics. London: Pinter Publishers.

 --- in press (b). New ways of meaning. In S. Efstathiadis, Ed. Selected Papers from the Ninth World Congress of Applied Linguistics. Thessaloniki: Applied Linguistics Association of Greece.

 --- in press ?. Towards probabilistic interpretations. In E. Ventola, Ed., Recent Systemic and Other Functional Views on Language. Berlin-New York: Mouton/deGruyter.

 Harrison, L. 1982. An overview of kinetic theory in developmental modeling. In S. Subtelny and P. Green, Eds., Developmental Order: Its Origin and Regulation. New York: Alan R. Liss

Hasan, R. 1989. The structure of a text. In Halliday, M.A.K. & Hasan, R. Language, Context, and Text. London: Oxford University Press.

Hasan, R. & Cloran, C. 1990. A sociolinguistic interpretation of everyday talk between mothers and children. In Halliday, M.A.K., John Gibbons, and Howard Nicholas, Eds., Learning, Keeping, and Using Language. Amsterdam: John Benjamins.

Hodge, R. & Kress, G. 1988. Social Semiotics. Ithaca: Cornell University Press.

Holling, C.S. 1986. The resilience of terrestrial ecosystems. In W.C. Clark & R.E. Munn, Eds., Sustainable Development of the Biosphere. Cambridge: Cambridge University Press.

Jackson, E. Atlee. 1989. Perspectives of Non-Linear Development. Cambridge: Cambridge University Press.

Lemke, J.L. 1983. Thematic analysis: systems, structures, and strategies. Recherches Semiotiques/Semiotic Inquiry (Toronto) 3: 159-187.

--- 1984. Semiotics and Education. Toronto: Victoria College/Toronto Semiotic Circle Monographs.

 --- 1985. Ideology, intertextuality, and the notion of register. In J.D. Benson and W.S. Greaves, Eds., Systemic Perspectives on Discourse, Vol.1. Norwood, NJ: Ablex Publishing.

 --- 1987. Strategic deployment of speech and action: a sociosemiotic analysis. In J. Evans & J. Deely, Eds., Semiotics 1983: Proceedings of the Semiotic Society of America. New York: University Press of America.

 --- 1988a. Discourses in conflict: Heteroglossia and text semantics. In J.D. Benson and W.S. Greaves, Eds., Systemic Functional Approaches to Discourse. Norwood, NJ: Ablex Publishing.

 --- 1988b. Text structure and text semantics. In R. Veltman & E. Steiner, Eds., Pragmatics, Discourse, and Text: Systemic Approaches. London: Pinter.

 --- 1988c. Towards a Social Semiotics of the Material Subject. In SASSC Working Papers 2 (1): 1-17. [Sydney: Sydney Association for Studies in Society and Culture]

 --- 1989. Semantics and social values. WORD 40(1-2): 37-50.

 --- 1990a. Talking Science: Language, Learning, and Values. Norwood, NJ: Ablex Publishing.

 --- 1990b. Technical discourse and technocratic ideology. In M. A. K. Halliday, John Gibbons, and Howard Nicholas, Eds., Learning, Keeping, and Using Language: Selected Papers from the 8 th AILA World Congress of Applied Linguistics. Vol. II. pp. 435-460. Amsterdam: John Benjamins. 1990.

 --- In press. Interpersonal meaning in discourse: Value orientations. In L. Ravelli and M. Davies, Eds., Selected Papers in Systemic Linguistics. London: Pinter.

 --- Forthcoming-a. Bodying pain and meaning.

 --- Forthcoming-b. Semogenesis and semodynamic development.

 Lovelock, J. 1989. The Ages of Gaia: A Biography of Our Living Earth. Oxford: Oxford University Press.

Martin, J.R. 1985. Process and text. In J.D. Benson and W.S. Greaves, Eds., Systemic Perspectives on Discourse, Vol.1. Norwood, NJ: Ablex Publishing.

Montagu, A. 1981. Growing Young. New York: McGraw-Hill.

Nesbitt, C. & Plum, G. 1988. Probabilities in a systemic grammar: the clause complex in English. In R.P. Fawcett & D.J. Young, Eds., New Developments in Systemic Linguistics, Vol.2. London: Pinter.

Odum, Howard T. 1983. Systems Ecology. New York: John Wiley.

Prigogine, I. 1980. From Being to Becoming: Time and Complexity in the Physical Sciences. New York: W.H. Freeman.

Prigogine, I. & Stengers, I. 1984. Order out of Chaos: Man’s New Dialogue with Nature. New York: Bantam Books.

Rosen, R. 1985. Anticipatory Systems. Oxford: Pergamon Press.

Salthe, S. 1985. Evolving Hierarchical Systems. New York: Columbia University Press.

--- 1989. Self-organization in hierarchically structured systems. Systems Research 6(3): 199-208.

Schneider, E. 1988. Thermodynamics, ecological succession, and natural selection. In B.H. Weber, D.J. Depew, and J.D. Smith, Eds., Entropy, Information, and Evolution. Cambridge, MA: MIT Press.

Thibault, P. 1986. Text, Discourse, and Context: A Social Semiotic Perspective. Toronto, Victoria University: Toronto Semiotic Circle Monographs.

--- 1989a. Semantic variation, social heteroglossia, intertextuality. Critical Studies 1 (2): 181-209.

 --- 1989b. Genres, social action, and pedagogy: a critical social semiotic account. Southern Review (Australia) 22 (3): 338-362.

 Thom, R. 1975. Structural Stability and Morphogenesis. Reading, MA: W.A. Benjamin.

Threadgold, T. 1986. Semiotics-Ideology-Language. In T. Threadgold, E.A.Grosz, Gunther Kress, & M.A.K. Halliday, Eds., Semiotics, Ideology, Language. Sydney: Sydney Association for Studies in Society and Culture.

--- 1989. Paradigms of Culture and Semiosis. In Walter A. Koch, Ed., Evolution of Culture. Bochum: Studienverlag Brockmeyer.

Ure, J. & Ellis, J. 1974. Register in descriptive linguistics and linguistic sociology. In O. Uribe-Villegas, Ed., Issues in Sociolinguistics. The Hague: Mouton.

Ventola, E. 1987. The Structure of Social Interaction: A Systemic Approach to the Semiotics of Service Encounters. London: Frances Pinter.

Wallace, A.F.C. 1970. Culture and Personality. Second Edition. New York: Random House.

Weber, B.H., Depew, D.J., & Smith, J.D., Eds., Entropy, Information, and Evolution. Cambridge, MA: MIT Press.