Go to Semiotic Review of Books Home Page
Go to SRB Highlights
Go to SRB Archives
This article appeared in Volume 5 (2) of The Semiotic Review of Books.
The notion of text constitutes the primary data on which much of linguistic and. more broadly, semiotic research is based. Yet. this notion has rarely been subjected to critical scrutiny. There may be culture-specific notions concerning what is and is not a text. In Western culture. and especially in its academic and pedagogical subcultures, the privileged model for text is that of writing. For such a model, text is a complete object. synoptically defined, which has determinate boundaries and which exists 'out of time'. The original model may be that of the sacred text. Text-as-writing induces us to think of text as a bounded and autonomous object which, in order to be understood or interpreted must be put in its context. In this model, there is a determinate sequence of wordings (lexicogrammatical forms) and these are taken to realize a stable and context independent semantics of the propositions which the forms encode. In this view, an autonomous and objective syntax encodes the literal meanings of their forms. However. this view which takes it that there is. ideally. a 'degree zero' of meaning which is independent of context. does not solve the problem as to what the text. so defined. is doing in context' (Silverstein. 1992 56). Unless this problem is resolved, the text-as-object is no more than a material entity - a cultural artefact - which has been produced in and through some culturally recognizable social activity-structure type. It is a product of the social semiotic processes which produced it.
In this view the meanings are somehow said to be 'in' the text and meaning is restricted to the supposedly context-independent ideational or propositional sort. The morphosyntactic forms of the text are simply a means for expressing something which is independent of these forms. Formal views of morphosyntax such as these entail a closed and positivistic conception of language form. Michael Reddy (1979) has aptly described this as the 'conduit' view of language. That is. syntactic forms are viewed as a sort of pipeline down which pre-existing and nonsemiotic propositional information. concepts. or contents are poured. Thus morphosyntactic forms are seen as containing their 'literal' (context-independent) meanings. which these same forms allow to be transmitted from a sender to a receiver. Language forms. in this view do not actively shape these meanings. Instead, they simply package them up and send them. Nevertheless. this view of meaning comes up against the problem of establishing the correspondences between the denotations expressed in morphosyntactic forms and the real world entities which these 'refer to'. This view may be said to have a theory of context. but it is one which is founded on some principle of indexical denotation. where the morphosyntactic forms index ('refer to') entities 'out there', by virtue of conventions which assign the same expression to the same referent (Austin 1989 (1952-3).
As a material artefact, text-as-object is always an incomplete record of the social activities that produced it. The (written) text is not a form of morse-code which simply stores and transmits pre-existing meanings without also transforming them. A written text is always a recontextualization from one system to another. The physical and material 'permanency' of the written text conceals this important and neglected fact. Thus, a written transcript of some original speech event is always a recontextualization of the semiotic modalities which were in operation in that event (speech, paralanguage, kinesics, gestures, and so on) transforming them into the semiotic modalities of the second (written) system. In this sense, a written text has the status of a textual record of the 'original' event: the meanings of the first system and their semiotic modalities are transformed in ways which are specific to the semiotic and material potential of the second system. Inevitably, this involves the loss of some semiotic modalities, the addition of others, and the transformation of still others. A further important aspect of this process is that the semiotic modalities of the second system - the written transcript, in our hypothetical example have "a significant potential for meaning production (and therefore for cultural transformation) in (their) own right" (Kress, 1990: 11). Text-as-record (e.g., a verbal transcript) is meant to be performed, i.e.,it is meant to be used and reenacted in some other social activity-type other than the one of which it is a record. This necessarily entails further recontextualizations of the text-as-record.
A text may also be a product of the social activity or performance in which it was made. This important distinction between text-as-product and text-as-record was first made by Jay Lemke (1984: 79-80: for further discussion see also Thibault, 1991: chaps. 3-4). A poem. for instance has this status as text-as-product. A textual product is a product, perhaps in a highly condensed and edited form of the semiotic modalities which constituted the social activity, Writing a Poem. Unlike text-as record text-as-product does not posit some relation of homology. or 'identity', between the social processes which enacted it - for example, the cultural activity of Writing a Poem - and its status as a textual product. All the same, there will be traces of these. however condensed and edited they may be in the final product. It is this lack of a homology which. perhaps. leads to the ideology of textual 'autonomy' for at least some classes of texts. Yet. text-as-product is also always a recontextualization in the sense I mentioned above. It is always a recontextualization of other social activities which are of types recognizable in the culture of the writer and readers of the text. The social activities which are recontextualized in this way may be indifferently linguistic or nonlinguistic classes of semiotic events.
The work of Fernando Poyatos (1983). for instance has indicated many of the ways in which this is true of narrative literature. A poem or a novel. say. is a recontextualization of the social activity types and the semiotic modalities of some socio-cultural group(s). This may entail the recontextualization of both verbal and nonverbal semiotic modalities (gestural. somatic. sartorial olfactory kinesic and so forth) into the meaning potential which is afforded by the semiotic modalities of the novel. Similarly a novel may be further recontextualized into the semiotic potential of film. In each case. the processes of recontextualization from one semiotic system to another always involve irreversible transformations of the prior semiotic system(s), as well as the activation of the semiotic potential which is specific to the new system(s). Considerations such as these point to the necessarily embodied nature of our participation as subjects and agents in social meaning making. The dominant theories of the sign in the Western tradition have, on the other hand worked assiduously to remove both the body and subjectivity from social semiosis (Thibault. forthcoming).
Now. texts. as I remarked earlier. are always created in and through their enactment in some social activity-structure. What is the relation of the text to the social activity-structure in and through which the text was enacted. performed. or produced? It is this question which has led speech act theorists. pragmaticians and others to invoke 'rules of use' or 'rules of contextualization' so as to interpret what the 'degree zero' of the context independent text is doing 'in context' (See Givon. 1989 73-5; Silverstein 1992:56: Thibault and Van Leeuwen forthcoming).
The difficulties of these approaches arise out of their recognition that context entails social phenomena. Context is considerably more than the given physical environment of the text and which the text simply 'reflects' by virtue of the correspondence relations between language forms and extralinguistic reality 'out there'. Immediately. the picture looks somewhat less tidy than was previously thought. Minimally social phenomena mean interaction between participants on socially specific occasions of discourse. This poses a problem for a view of language which is founded on the rationalistic and asocial cognitive competence of individual speakers. How. in such a view might context be handled without, however jeopardising the basic assumptions about language and its relation to an individual-centred rationality? Typically, the 'problem' of the social-interactional dimension of language is removed from language altogether and exported to a 'pragmatic' level of analysis (see Thibault and Van Leeuwen. forthcoming). Thus, pragmatic 'rules of use' or 'maxims of interpretation' of the Gricean sort are pressed into service to explain what the text is doing 'in context'. These tend closely to reflect local, agent centred folk-ideological accounts of language as a means or instrument for attaining individual goals.
The notion of text which is at issue here is somewhat different in status from that of the object text which I referred to earlier. Here. we have a second view of text, which I shall refer to as the semiotic action text. The semiotic action text is dynamic and processual, rather than synoptic and object-like. It is the text which unfolds in the real-time production of some social activity structure. This may also entail the production of an object-text, or its reading, re-enactment, or performance as a contingently new meaning text in some social activity-structure-type other than the one in which it was produced. Both of these cases - reading or re-enacting - stand in an internal and homological relation to the social activity structure in and through which the text is either produced or read. However, the type and degree of this homology differs in the two cases. Doubtless, this is not reducible to a simple two-way distinction, and a complex. multidimensional topological space would probably provide the best means for describing the various types of homology which occur. Notice, too, that I have used the term homology, rather than 'identity'. The semiotic action text is not identical to or reducible to the social activity-structure in and through which the former was produced and/or enacted. Further. this is so along all of the various dimensions of. say, linguistic meaning which are constituted by the given social activity.
How can we show that there is an internal and homological relation between the text and the social activity? What are the implications of this? If one starts with the view that meaning is the correspondence relations between autonomous morphosyntactic forms and some features of the context. where context is defined in positivistic and nonsemiotic terms as observable entities, states of affairs, and so forth 'out there', then the relations between the interpretative maxims or rules of use and syntactic form is an external, contingent, and essentially ad hoc one. But if we can show that the internal organization of language and text is multifunctional, i.e. that semantics involves more kinds of meaning relations than those of the referring-and predicating sort, - then a very different picture of this relation comes into view. We can argue, for instance. that other sorts of meanings than the propositional or ideational kind are a constitutive sort of the intrinsic organization of language, and other semiotic forms. For example, a textual record such as a written transcription of some prior socio-discursive event can tell us a great deal about the semiotic structure of that event. This is so because the text (the transcription) is itself a semiotic structure which realizes sets of meaning relations, which can in their turn tell the reader a good deal about the social activity-structure. of which our text is a transcription. It cannot tell us everything about that original social activity structural because, as discussed above, text-as record always entails the loss and/or transformation of some aspects of the original meaning potential. as well as the addition of new potential. according to the semiotic modalities in which the textual record itself is made. This is no less true of audio and visual recordings than it is of verbal transcriptions. These are all forms of text-as-record. What kinds of information can the textual record tell us? What kinds are missing or transformed? And how are they transformed?
The kinds of meaning relations which will tell the reader or analyst something about the original socio-discursive event have been systematized below. A text will evidence, or is a trace of, the following four types of meaning relations:
Now, in our hypothetical example of a written transcript of some original speech event, we can see how the textual record construes the original socio-discursive event along these four analytically reconstructible dimensions of social semiosis. We can point to the text as evidence of some prior socio-(discursive event which was instantistod in some time and place. But does the case of the literary text contradict this? Only apparently, i e.. if one assumes that the literary text is 'autonomous' to start with. However, a poem or a novel is already a recontextualization into other semiotic modalities of socio-discursive events, and these are recognizable as types of social activity-structures in the culture of writers and readers of such texts. The poem or novel may or may not stand in quite the same relation to a specific. prior event as does our verbal transcription. Nevertheless, it is a recontextualization, into other semiotic modalities, of activity structure-types which are socially and culturally specific. The point here is that the four dimensions of semiosis which I have identified above do not simply reflect a preexisting reality which stands in an isomorphic relation to our transcription.
Nor is context itself a pre-existing reality which contains the text. Instead, texts, are constitutive components of their social activity-structures. They are both constitutive of and constituted by the social activity which enacts and/or performs them. They are analytically, but not constitutively. separable from these. There is, then, a two-way and dialectical relationship between the two levels of analysis - between text and social activity. The nature of this dually constitutive and constituted relation would appear to differ from type to type. A given text may also combine both aspects of text-as-record and text-as-product. But the basic principle still holds. Thus, the conversational text which unfolds during the course of the social activity of Eating a Meal Together may have very little to do with the no less social activity of Eating a Meal Together. It may have, of course, as when speakers, in the course of their conversation, index features of the social activity, Eating a Meal Together. But this is not necessarily so: the conversational text may also be on quite different cultural topics. To be sure, the dynamic unfolding of the text responds to different phases in the activity which is taking place, when it takes place, but this does not necessarily mean that, in the conversational text itself. these directly respond to any particular socio-discursive feature of the social activity of eating a meal together.
Thus, a verbal transcription which confined itself to the semiotic modalities of the conversational text in the narrow, linguistic, sense would, in effect delete that whole dimension of semiosis which I have glossed as Eating a Meal Together. The conversational text is enacted in and through are organized in ways which are not dissimilar to this social activity-type. A transcription which is semiotically informative must find ways of transcribing the composite semiotic action text, in all of its semiotic modalities.
It is quite wrong in my view, to assume that the social activity of Eating a Meal Together is the 'context of' the conversational text in any simple, external way. Yet, it is precisely this kind of reasoning which has tended to prevail in our tendecy to separate 'text' from 'context'. Typically, in this way of doing things, 'context' becomes all that is left over once the verbal text has been transcribed as a textual record, or otherwise accounted for. Context hten becomes a set of ad hoc rules or maxims for interpreting a linguistic semiotic whose systematic codeployment with other (non-linguistic) semiotic resource systems has been, if not ignored, than at least trivialized and relegated to a set of extrasemiotic 'prelocutionary' or other effects.
This brings us to a critically important issue. Instead of a view which seeks to add context to text, or to put the text 'in context', we need a semiotic description which is able to provide a record of description which is able to provide a record of the composite semiotic action text - the socio-discursive event - in and through which texts are produced and interpreted. This brings us to two key dimensions of semiosis which have received all too little attention. These may be summarized as follows:
Language has usually been seen as only weakly cross-coupled with either or both of the two dimensions referred to here, as Lemke (1991a) has argued. As a consequence, we have the theoretical fiction that language is (1) form, but not substance; and (2) that it is an autonomous system, not systematically related to other modalities of social semiosis. Certainly. Hjelmslev (1954: 173) recognized that substance is "semiotically formed", yet this insight has not been systematically developed.
The importance of point (1) is that it highlights the way in which a social activity-structure type is constructed in and through the joint deployment. by the participants to it, of the semiotic and material resources in order to enact a regular, repeatable activity-structure of a given type. The semiotic-discursive resources are deployed in ways which coordinate and entrain physical material states and processes. The further point to make here is that social activity-structures of a specifiable type are themselves semiotic-material performances. Just like the texts which they produce or enact in determinate ways, social activity-structures themselves deploy semiotic resource systems in regular and repeatable ways. These may include all the semiotic resource systems which may not be included in a verbal transcription, or only marginally so Thus, actional. paralinguistic. kinesic, somatic, and other semiotic modalities are seen as merely 'contextualizing' or 'disambiguating' the linguistic text. which is seen as having primary status. This now brings us to the importance of point (2) above.
I have said that language is strongly cross coupled with the other, nonverbal semiotic modalities. In other words, there are strong intersemiotic cross-couplings. This has a number of important consequences. First, language has both co-evolved with and is always co-deployed with other semiotic resource systems (Lemke, 1991a). There are no exceptions to this, including the seemingly emblematic case of the written text. Written language always simultaneously codeploys both visual and linguistic semiotic modalities in the making of texts (Kress and Van Leeuwen, 199C). Secondly, the four general semiotic functions, or modes of meaning making, which I mentioned earlier, occur in all of the various semiotic resource systems. All semiotic resource systems are internally organized in ways which reflect these four dimensions of semiosis. ln other words, no act of semiosis occurs without these four dimensions being simultaneously activated. Recent work on the visual semiotic by Kress and Van Leeuwen (1990),on music by Van Leeuwen (1991), and on sign language (Johnston, 1992) provide impressive theoretical arguements and analyses which suggest the trans-semiotic functions mentioned above. The earlier work of Scheflen (1973) also provides interesting insights into the ways in which the specifically somatic semiotic modalities are organized in ways which are not dissimilar to those proposed earlier. Perhaps the main limitation of much of this earlier research in this area, inspired by the poineering work of Ray Birdwhistell, was the failure to understand that these semiotic modalities are internally shaped and organized by the specifically 'semantic' level of their organization. The tendency was to describe in minute detail the level of expression form of these semiotic modalities, but without seeing in any systematic way how the content form of these semiotic modalities is organized in relation to their expression form. That position has begun to change in recent years (Johnston, 1992, Cuxac, 1993).
The four semiotic functions I mentioned above are a generalization from and a futher development of Michael Halliday's semantically oriented theory of language form and function. In that theory, grammar is niether formal or autonomous. Instead, the grammar itself is internally organized in ways which reflect the four dimensions of semiosis discussed above. Grammar is a system of choices - a semiotic resource - for making meanings along these four dimension of semiosis simultaneously. A given choice always realizes a selection of meaningful options from each of these four systems. Further, these are simultaneously configured in language and other semiotic forms, and in ways which leave a distinctive trace in form itself. Halliday (1979) has referred to these as the experiential, logicalm interpersonal, and textual modes of meaning, which correspond to the four dimensions of semiosis specified above. Halliday claims that it is possible to specify the meaning potential of a linguistic form, or that of an entire text, in terms of the semnatic resources hatt it simultaneously deploys along thse four dimensions of semiosis.
These are intrinsic to language form, and this is an important difference with respect to the earlier functional theories of, say, Buhler and Jakobson, which remained extrinsic in the sense of that the functions these scholars postualted were not related to the internal design features of the language in any systematic way (Martin, 1991). The point is that language form is intrinsically organized in the way it is as a consequence of the multifunctional nature of semiosis itself.
There is an internal semantic redundancy among the various modes of semiosis. These are, as I have said, simultaneously co-deployed in the making of texts and social activity-structures. Further, these are co-deployed in patterned ways it is possible to specify selection probabilities and joint probability distributions of selections from different systems, or, indeed, from the same system at different points in the same text, as well as those conditional selection probabilities which link what is going on in one part of the same text to some other part of the same text, or its intertexts (Lemke, 1991b). It is only when all four of these semiotic modes are co-deployed that semiosis occurs. This also means that the internal design features of all semiotic forms are always cross coupled with both the pre-semiotic physical material domain and with other semiotic modalities. This is so along all four of the dimensions I have mentioned here.
But what of 'context'? If language and the other semiotic resource systems are strongly cross coupled in social semiosis, as well as being strongly cross-coupled with physical-material processes, then where does 'context' fit in? The answers to these questions require a radical reorientation of our theoretical thinking concerning, in the first instance, the nature and status of semiotic form itself. According to the dominant tradition, this is simply the vehicle or 'conduit' for the coding and transmission of information, or content, which pre-exists these forms, and the sole purpose of which is to package this informational or conceptual content for sending off to a designated receiver, or decoder. But if, alternatively, the internal design features of semiotic forms are organized along the lines discussed above, then we can say that signs construct and construe what lies 'outside' them along all four of these dimensions. The relations between the internal design features of the forms and that which they construe in the 'outside' world is, to use the language of quantum physics, one of complementarity. Semiotic forms do not, in this conception, simply 'stand for', reflect, or represent a pre-given reality 'out there'. Both Saussure and Whorf had understood this, and in ways which mainstream linguistics have failed to heed ever since, i.e., that the intrinsic design features the language - its grammar - is categorical (see also Langacker, 1987; Petitot-Cocorda, 1990 (1985)). Both Saussure and Whorf, however, and in their different ways, limited their thinking to the experiential, or propositional, dimension of semiosis.
My point is that it is, to use the Saussurean terminology, the internal value-producing potential of semiotic forms which determines, in semiosis, that which is relevant in the 'outside'. It does so along the four dimensions already described, by making that which is relevant emerge from a presemiotic 'background' of physical-material states and processes. Thus, physical-material states and processes are made salient and meaningful in semiosis in and through the cross-coupling patterns they enter into with the various modes of semiosis which are built into the very internal design of semiotic form itself. This has nothing to do with the slicing up and classifying of 'external' reality by some pre-existing grid of conceptual or other categories. Meaning making is the process of cross-coupling the multifunctional potential of semiotic forms with their 'outside'. This happens in ways which are never fixed or wholly determinate. Rather, they are dynamically changing in the real-time production of discourse and in ways which ensure that the nature of this cross-coupling always retains a degree of undecidability or ineffability. This is so because there is not, as in the folk-linguistic theory of reference, a correspondence value between a given linguistic feature, typically the grammatical Subject in the orthographic sentence, and a given entity in the real world 'out there'. It was precisely this folk-ideology of reference which Whorf deconstructed so insightfully in his subtle analyses of the dialectic between language form and function and the linguistic praxis of a given community.
The cross-coupling of the semiotic-discursive and the physical-material domains does not occur on a one-to-one or correlational basis. Instead, it is the globally co-patterned nature of the internal redundancies among the various dimensions of semiosis which gets cross-coupled with the 'outside'. The principle of complementarity to which I alluded earlier means that the internal design features of semiotic forms are also structured by that with which they are cross coupled: that which is construed by these forms is also incorporated into and shapes their internal organization. This is so because the linguistic and other semiotic modalities have both co-evolved and are co-deployed with physical-material processes and with each other. This does not, however, entail any reduction of the one to the other. The multifunctional and globally copatterned nature of the various dimensions of semiotic form means that these always contextualize one another. They also contextualize the 'outside' at the same time that this contextualizes the 'inside'.
Language and the other semiotic resource systems do not then provide a simple window on a given reality. Quantum mechanics has taught us to give up the notion that an object is localizable in a system of well-defined mechanical parameters (Prigogine and Stengers, 1985: 224). It is the interaction between the globally organized system of relations and the 'outside' which is relevant. It is not the case that semiotic form X and real-world entity Y stand in a definite relation to each other. The globally co-patterned nature of semiotic forms. in the real-time dynamics of discourse production. means that different weightings of the co-selection probabilities and the joint distributions of the various systems involved will foreground or background different possible copatterninqs of these and in ways which demonstrate the multiplicity of the possible construals of the Real.
These reflections suggest that if the place of language in semiosis needs to be put off-centre, then this also requires a radical rethinking not only of language in relation to the other semiotic resource systems with which language is codeployed but also of the internal nature of all semiotic forms both in relation to each other and in relation to the processes of social semiosis which these forms help to constitute.
Austin, J.L. 1989 (1952-3). 'How to talk'. In J.O. Umson and G.J. Warnock (Eds.). Philosophical Papers. Third Edition, pp. 134-53. Oxford and New York: Oxford University Press.
Cuxac,Christian.1993.'Iconicite dans Les Langues des Signes'. Faits de Langues 1, March 1993, pp.47-56.
Givon, T.1989. Mind, Code and Context, Essays in Pragmatics. London and Hillsdale, NJ: Lawrence Erlbaum.
Halliday, M.A.K. 1979. 'Modes of meaning and modes of expression: types of grammatical structure, and their determination by different semantic functions'. In D.J. Allerton, Edward Carney and David Holdcrofy (Eds.). Function and Context in Linguistic Analysis. A Festschrift for William Haas, pp. 57-79. Cambridge: Cambridge University Press.
Hjelmslev, Louis. 1954. 'La stratification langage'. Word 10, 2-3, pp.163-88.
Johnston, Trevor, 1992. 'The realization of the linguistic metafunctions in a Sign Language'. Social Semiotics 2,1, pp.1-43.
Kress, Gunther. 1990. 'Media literacy as cultural technology in the age of transcultural media'. Keynote address to the colloquy on New Directions in Media Education, Toulouse.
Kress Gunther and Theo van Leeuwen. 1990. Reading Images. Geelong, Victoria (Australia): Deakin University Press.
Langacker, Ronald W. 1987. Foundations of Cognitive Grammar. Vol. 1. Theoretical prerequisites. Stanford, CA: Stanford University Press.
Lemke, Jay L. 1984. 'Action, context, and meaning'. In Semiotics and Education, pp. 6393. Victoria University, Toronto: Toronto Semiotic Circle Monographs, Working Papers and Prepublications, No.2.
- 1991a. 'Discourse, dynamics, and social change'. To appear in M.A.K. (Issue Ed.). Language as Cultural Dynamics (Special issue of Cultural Dynamics, Ed. Rik Pinxten).
- 1991 b. 'Text production and dynamic text semantics'. In Eija Ventola. Ed. Functional and Systemic Linguistics. Approaches and uses, pp. 23-38. Mouton de Gruyter: Berlin and New York.
Martin J.R. 1991. 'Intrinsic functionality: implications for contextual theory'. Social Semiotics 1, 1, pp. 99-162.
Petitot-Cocorda, Jean. 1990 (1985). Morfogenesi del senso. Per uno schematismo della struttura. Italian trans. Marcello Castellana, Mario Jacquemet, and Francesco Marsciani. Milan: Bompiani.
Poyatos, Fernando, 1983. New Perspectives in nonverbal Communication. Studies in cultural anthropology, social psychology, linguistics, literature and semiotics. Oxford and New York: Pergamon Press.
Prigogine, llya and Isabelle Stengers, 1984. Order Out of Chaos. Man's new dialogue with nature. London: Fontana.
Reddy Michael. 1979. 'The conduit metaphor- a case of frame conflict in our language about language'. In Andrew Ortony (Ed.) Metaphor and Thought, pp. 284-324. London and New York: Cambridge University Press.
Scheflen, Albert E. 1973. How Behavior Means. New York: Gordon and Breach.
Silverstein Michael. 1992. 'The indeterminacy of contextualization: when is enough enough?'. In Peter Auer and Aldo di Luzio (Eds.). The Contextualization of Language, pp. 55-76. Amsterdam/Philadelphia: John Benjamins.
Thibault, Paul J. 1991. Social Semiotics as Praxis. Minneapolis and Oxford: University of Minnesota Press.
- Forthcoming. Re-reading Saussure. London: Routledge.
- and Van Leeuwen, Theo. Forthcoming. 'Grammar. societv. and the speech act: renewina the connections'. Mimeo: Dipartimento di Linalgistica. Universita di Padova (Italy).
Van Leeuwen. Theo. 1991. 'The sociosemiotics of easy listening music'. Social Semiotics 1, 1, pp.67-80.
Paul J. Thibault is Professore associato in English Language and Linguistics, Universita di Padova, Italy. Born in Newcastle, Australia, he studied linguistics and English language and literature at the University of Newcastle before completing his Ph.D. with Michael Halliday at the University of Syndney. He has taught at Murdoch University, and the Universities of Sydney, Verona, and Bologna. Research interests include functional grammar and semantics, discourse analysis, genre theory, semiotics, literary and social theory and, increasingly, child language, educational linguistics and literacy. He has published a monograph, Text, Discourse, and Context: a social semiotic perspective, in the Toronto Semiotic Circle series (1986), articles, book chapters and a book, Social Semiotics as Praxis, (Minnesota, 1991). He is currently completing a book for Routledge entitled Rereading Saussure, as well as working on two others on childrens' play and the semantics and grammar of interpersonal meaning.