Image/Text
Paul Filmer
I seemed to learn . . .
That what we see of forms and images
Which float along on minds, and what we feel
Of active or recognizable thought,
Prospectiveness, or intellect, or will,
Not only is not worthy to be deemed
Our being, to be prized as what we are,
But is the very littleness of life.
Such consciousness I deem but accidents,
Relapses from the one interior life
That lives in all things, sacred from the touch
Of that false secondary power by which
In weakness we create distinctions, then
Believe that all our puny boundaries are things
Which we perceive and not which we have made.
William Wordsworth
IMAGE/TEXT AND SOCIOLOGICAL DISCOURSE
The sociological dichotomy between image and text is one which has emerged only recently, though Wordsworth reminds us that it is a much more venerable preoccupation of Enlightenment thought about the dichotomous tension between culture and nature (see chapter 'Culture/ Nature'). As a concern for sociology, however, it is largely a result of the
CORE SOCIOLOGICAL DICHOTOMIES
importance of structuralism in twentieth-century social and cultural theories. This relates the dichotomy, both directly and indirectly, to the significance for sociology of concepts and theories of language as a social phenomenon that expresses consciousness intersubjectively. Yet sociology was marked by a reluctance to address language as a topic until the middle of the twentieth century, largely because of the desire of sociologists working within the dominant normative schools to preserve a positivist methodology for a scientific sociology. This involved a commitment to a correspondence version of truth and an attendant version of language as a neutral, correspondential medium of communication and expression. Both versions served well enough the epistemological ambitions for sociology as a science, but neither was able adequately to accommodate the reflexively experiential character of the social world, and the extent to which it was grounded by and for its actors in the negotiated truths of common sense which are established in and expressed through language. Whilst the potential for considering this linguistically constituted character of the social world was available conceptually in the formulation of the individual as social actor (see chapter "Theory/Practice') and methodologically in the formulation of purposive social action as agency (see chapter 'Structure/ Agency'), the philosophical positivism underlying these conceptions denied them the possibility of being put into operation in terms of the social significance of language as the basis of sociological explanation. This gave the image/text dichotomy a history within sociology during this period on other terms, and in related forms: for example, as a dichotomous tension between the differentiated referents and manifestations of symbol and sign, where the sign establishes a concrete and specific (textual) realization of the imaginative possibilities represented by the symbol from which it is derived. It can be seen as implicated also in the idealism/materialism dichotomy, in the sense that a text may stand for an empirically specific, and therefore materially constraining, version of an imagined ideal (see chapter 'Idealism/Materialism'). Most significantly, perhaps, there is a clear sense of the dichotomy which lurks in the discursive sociological differentiation between theoretical images of society as a whole and the constructed texts of empirical sociological data.
The advent and consolidation of structuralist theories of language as a system of signs through the first two-thirds of the twentieth century has turned sociology back from its reluctance to address language and towards a concern with its essentially social character. This is not to suggest, however, a sociologistic account of language. On the contrary, it is clear that the origins of human society and language are dialectically interdependent and equally prehistoric. They are not empirically available for historically specific explanation. Instead, sociology addresses language theoretically and conceptually in terms of a phenomenology of the linguistic sign, though it is able to undertake detailed empirical inquiries into the embedded practices of its conventions of usage and the mundane meanings
300
IMAGE/TEXT
implied in them. Indeed, these conventions and meanings enable some empirical sociologists to claim variously, and quite coherently, that language is a social fact and can be treated functionally as if it is an interactionally natural phenomenon (Giglioli 1972; Boden and Zimmerman 1991).
Image/text relations have become a focus of central contemporary sociological interest as a result of the instatement of language as a topic of serious sociological concern and, following from this, the important interrelations which sociology has developed with other disciplines during the second half of the twentieth century, through adopting in common structuralist methodological approaches to the analysis of contemporary culture (see chapters 'Culture/Nature', 'Modernity/Postmodernity' and 'High/Mass')- The most important of these interrelations have been with cultural studies (Turner 1990; Davies 1993) and literary theory (Eagleton 1976; 1983) and have led to further involvement with contemporary psychoanalytical theory and with critical feminist thought (Milner 1994; see also chapter 'Theory/Practice')- These developments have had significant implications also for the sociology of art (see, for example, Wolff 1981; 1983; Filmer 1998), where they have generated attempts to extend the range of this subdisciplinary area beyond attempts to explain works of visual and literary art in the rather limiting terms of contextually determined reflections of their social worlds, into critical analyses of the reflexive relations between artistic practices and the social structures in which they are generated and to which they refer (see Natanson 1970; Blum et al. 1974; Bryson 1981). In structuralist terms, the result of these developments has been the elaboration of a series of methodological strategies which involve the treatment of social practices as if they are texts which articulate prior imagistic conceptions and projections of orders of relations (structures) within and between human groups which already exist, but of which members of the groups are not yet conscious, or which may yet be able to be brought into existence. These implicit and imaginary structures are complemented by possible intended meanings (senses) that members of the groups already share or might come to hold in common and which also require articulation in ways that may be treated as texts.
IMAGE, TEXT AND SOCIAL REALITY: THE DIVERSITY OF IMAGE/TEXT RELATIONS
The methodologically strategic, structuralist relation between image and text is quite abstract, and can be formulated clearly as a type of relation between substantive, empirical absence and presence, with several variants such as: image is absent, text is present; an image exists in consciousness but not as physical substance; a text is a substantive, empirical object; and
301
CORE SOCIOLOGICAL DICHOTOMIES
so on. As Leppert notes, 'what images represent may otherwise not exist in "reality" and may instead be confined to the realm of imagination, wish, desire, dream, or fantasy' (1996: 3), a point made more polemically by Boorstin's (1962) formulation of image as pseudo-event. Roland Barthes (1915-80) offers unequivocally as 'the definition of the image, any image: that from which I am excluded' (1978: 132) and, in writing of privacy, insists that it is 'the absolutely precious, inalienable site where my image is free (free to abolish itself)' (1984: 98), thus offering a further example of the image/text relation as that between private and public. Barthes offers these and other (e.g. 1977) formulations in relation to the social and aesthetic practices that constitute photography, one of the specific sites on which the dichotomous tension between image and text has been addressed recurrently. This dichotomy becomes the more clear in the case of photography because the concept of image is addressed in tension not only with the text, but with a sense that may be seen as extending beyond the text, to a reality which is external to it, and its photographic representation.
This tension, as Sontag (1978) points out, is one which, like the discipline of sociology itself, is a consequence of the Enlightenment. The advances in humanist and scientific thought of the nineteenth century were greeted with the expectation that they would bring within reach ways of representing the real that would be freed from the interpretive mediation of an image. The techniques of photography and film in particular were seen as supporting the progress of scientific research through their capacity to represent things, in full detail, precisely as science had discovered they were constituted and events exactly as they occurred. From the mid nineteenth century, an aesthetic of realism developed which argued that the tasks of visual art and literature should be directed to the work of 'mirroring nature' - a task which, though realist artists found it impossible to accomplish, nevertheless continues to sustain a critical rhetorical force (Foster 1980).
Indeed, this tension between the aim of 'mirroring nature' through linking the technology of visual reproduction with the aesthetics of realism relates to a cognate tension (Benjamin 1969; 1979; Gombrich 1956; 1982) between photography and visual art which elaborates the image/text dichotomy still further. A figurative visual art work - for example, a painted portrait or a landscape painting - may be interpreted as presenting an image of the artist's and /or viewer's perception of what it represents to them and as such, to constitute a text (Phillipson 1985). As a text, it can be compared to a photographic text of the same subject in terms of a criterion of verisimilitude, for example. Each text, moreover, can be transformed in each medium of representation to an image of the other: a photograph can be made of the painting, and a painting of the photograph. The latter transformation is a particularly complicated one because of its history as a conventional feature of artistic practice: Courbet, Manet, Monet, Degas and Cezanne, for example, regularly used photographs of their subjects as aides-mémoires for their paintings (Macdonald 1979). Far from defining the image
302
IMAGE/TEXT
more precisely, the continuing development of increasingly sophisticated technologies of visual reproduction has served, with the most creative of ironies, to make more complex the relations between image and text. As Sontag has noted, 'the credence that could no longer be given to realities understood in the form of images was now being given to realities understood to be images, illusions' (1978: 153). Whereas the post-Enlightenment expectation had been that scientific knowledge would come to supplant the 'magic' of imagistic representation, it came instead to depend upon it as its necessary antithesis (Warnock 1976). Moreover, by the middle of the twentieth century, and notwithstanding the considerable achievements of scientific research and technology, it had been clearly established that the correspondential truth claims of scientific positivism, which depended upon proof, could not themselves be proved absolutely (see, for example, Popper 1962; Kuhn 1970). However, the hermeneutic necessity for interpretation of claims to validity that this produced, especially in the social sciences, had already begun to be addressed much earlier, during the nineteenth century, even as the claims of aesthetic realism began to be announced (Hirsch 1967).
Marx's critical analysis of the culture of capitalism, though the materialist character of so much of his argument means inevitably that it comes to be committed almost entirely to realist aesthetics, nevertheless offers an early example of the significance of the concept of image for sociological thought. He is critical, for example, of the process of what he terms reification, whereby, for example, the interests of a particular dominant social group, the beliefs and values in terms of which they legitimate their social position and the social structure in which they are located, are all treated as an inevitable, unchangeable, even a 'natural' set of institutional arrangements. Similarly, he criticizes the tendency of capitalist culture towards fetishization of commodities, where material objects, like monetary coins, are treated as having intrinsic value rather than representing the value of the goods and services whose exchange they facilitate. In both instances, the practice which is criticized is one which substitutes an image, an illusion, for a reality. Indeed, Marx formulates ideology quite specifically in the imagistic terms of a camera obscura, in which individuals and their relations appear upside down (Mitchell 1980; 1986).
These criticisms invoke also a further sense of image which is sociologically important - that of an ontology of the human individual as a member of society (Lyon 1983). All general sociological theories are based upon an implicit or explicit formulation of the essentially social character of human being which informs their major arguments. In Marx's case, both of the imagistic fallacies - of reification and fetishization of commodities -which he sees as perpetrated by the dominant interest groups in capitalist societies are generated by, and in turn contribute to, the substantiation of his ontological concept of the human being as Homo laborens - the person who works and, by doing so, produces and reproduces the material conditions required to sustain human existence. Just as human individuals
303
CORE SOCIOLOGICAL DICHOTOMIES
reproduce themselves and their material conditions, so they are the principal and active agents in the production and reproduction of their social world and its constitutive political, economic and legal institutions and practices. In this sense, Marx is arguing that human beings reproduce themselves in their own image. The distinctive character of Marx's version of human ontology can be seen by contrasting it with that of his mentor, Georg Friedrich Hegel (1770-1831), for example, whom he later claims to have 'turned on his head'. Like many other philosophers, Hegel argued that the essential character of the human individual was that of Homo sapiens -the person who is wise and for whom thought, therefore, determines action. Marx's inversion of Hegel occurs with his contradictory proposition that it is the material conditions of human existence that determine consciousness. But both theorists are working with a clearly formulated and distinctive image of the human individual, in terms of which they develop their account of social reality. There is a sense, thus, in terms of which the image/text dichotomy parallels closely that between theory and practice.
This can be seen still more clearly if we consider Durkheim's contribution to the sociological formulation of image as a concept. Whilst Durkheim does not address image specifically, it is clear, as Smith's (1995) discussion implies, that his conceptions both of social fact and of society itself are imagistic and, in the Barthesian terms noted above, substantive absences rather than the empirical presences that his (Durkheim 1938) commitment to positivism would appear to require (Morrison 1995). Yet Durkheim implies a more significant sociological concept of image in his discussions of the relations between symbolism and society. These occur in his late works on Primitive Classification (Durkheim and Mauss 1963) and in The Elementary Forms of the Religious Life (1915), in the first of which he argues sociologistically (see Tiryakian 1962) that the abstract, symbolic categories in terms of which human societies classify their environments and experiences, such as time and space, are generated from the categories in terms of which they classify the structures of their own social relations. Thus space is conceptualized in terms of the segmental unity of a society of clans, each of which is composed of a network of extended families; and time is structured according to the principal events of the societal calendar of continuity, change and renewal. Each symbolic category is represented as an image of the social structures and processes which generate it. This is complemented in the later work on The Elementary Forms of the Religious Life by the argument that the sacred objects and ritual practices of religious worship are images of the society in which they are established. They are generated by the society as a reflexive representation of itself, on the grounds that the abstract idea of society can in this way be made cogni-tively accessible and understandable to its individual members as an emergent whole (see chapter 'Theory/Practice'). Thus the divine image itself (God) is seen by Durkheim as a projection of the collective consciousness, an imagistic representation of society in the form of a symbol which is
304
IMAGE/TEXT
socially generated. This amounts to a sociologistic inversion of the Judaeo-Christian myth that society is created in the divine image, since Durkheim has argued not only that society has created the image of the divine to represent itself, but also that it has generated the symbols with which to do so (Filmer 1977a). Thus, what he and Mauss term the 'total social fact' of emergent, collective consciousness is a symbolic reality, constituted in terms of a reflexive image of the structure of the social relations in which it inheres. The essential human capacity to symbolize which is implied here, and argued in Primitive Classification, suggests an ontological conception of the human individual in Durkheim's late work, as one who symbolizes. In the sense in which Durkheim conceptualizes them, moreover, symbols articulate the relation between image and text by representing the absent image as a substantive presence in some categorical or classificatory form (Lukes 1973). This form is conceptualized linguistically, and the symbol thus both articulates and represents a linguistic structuralist relation between image and text.
As a sociological phenomenon, as Leppert points out, 'any image literally exists as an object within the world that it in one way or another engages. When we look at images . . . what we see is the product of human consciousness, itself part and parcel of culture and history . . . images are constructed for the purpose of performing some function within a given sociocultural matrix' (1996: 3). It is to the realization of the possibility of performance of this function that the transformation of image into text is of crucial importance. In order to understand how that process of transformation occurs, it is necessary to consider the ways in which language mediates between image and text, and thus to consider some aspects of the textuality of language itself.
IMAGE, TEXT AND LANGUAGE: THE CONTINUITY OF IMAGE/ TEXT RELATIONS
The relations between image and text are structured invariably by language, and are therefore both semiotic and social. They are semiotic because they are constituted in and as signs which textualize the image; and they are social in so far as the textualization of the image makes it meaningful to the human collectivity which authors and shares the system of signs. The semiotic articulation of the image makes it analysable both on its own terms and in terms of the social conventions which make it meaningful. But this very accessibility means that at the very point at which an image is recognizable as a phenomenon in itself, it is made available for analysis: indeed, it is that availability which makes possible the constitution of the image as a phenomenon itself. Without its re-presentation in textual form, the image remains the 'absence' to which Barthes condemns it. The
305
CORE SOCIOLOGICAL DICHOTOMIES
process of analysis begins, thus, with the (re-)production of the image as a text (Barthes 1967a). This is a routine, micro-interactional and interpretive social process and, as such, is deeply embedded in the mundane practices of everyday social life. Like all such practices, however, its routine character belies the complexity which becomes apparent when attempts are made to analyse its structure. That depth of embeddedness makes the semiotics of image/text relations recondite for analysis. But what gives especial complexity to sociological analysis of relations between image and text is that the textualization of an image which makes it analysable involves also a transformational intervention into the conventions of language itself. In translating an image into a text, the language of the text is extended beyond an assembly of potential meanings established by conventions of usage, which are already and relatively unproblematically known to readers, into a quite new and thus potentially problematic configuration of meanings which may not be interpretable in terms of existing linguistic conventions. Airman refers to this as 'doubly articulated language' (1981: 40), a concept which makes it possible to differentiate between what he has termed the linguistic and the textual meaning of a sign. The established, conventional meanings of linguistic signs constitute a primary modelling system which makes sense of the mundane world and enables it routinely to be taken for granted, much as Schutz (1962) proposes when referring to the meanings of everyday words as first-order typifications. Texts contain linguistic signs which carry everyday meanings; but Altman suggests 'that the text's most radical and disconcerting method of making meaning is that which derives from its status as secondary modelling system. The experience of reading draws its special nature from the necessity of learning a new language in the course of reading . . . one must learn the language and construe a message during the same reception' (1981: 42).
Airman's suggestion recalls other formulations of the double articulation of language. He uses himself (1981: 40) the primary morphemic/phonemic relation of general linguistics between sign and sound, the relations between which, in the primary modelling system of everyday language, are arbitrary in their conventionality and are learnt and relearnt, apparently routinely and straightforwardly, as a mundane pattern of common-sense usage. This echoes Saussure's (1959) distinction between signifier and signified as the constitutive relations of the sign, which Saussure also proposes are conventional and arbitrary in character. For Saussure, the primary modelling system of language, termed la langue, is different from that of Altman in constituting a basic grammatical system which is generative of the langage of everyday life, equivalent to Airman's conception of sound. But again, both formulations hold in common that the conventional relations between signs and the phenomena to which they refer are arbitrary.
But this double articulation of language is expressed, in these formulations, at only a primary level. Schutz (1962) indicates that there is a
306
IMAGE/TEXT
double articulation between this and a higher level when he complements his concept of the first-order typifications which constitute the taken-for-granted meanings of everyday words and expressions with a concept of the more formal expressions which he terms second-order typifications. The relations between the signs which carry these higher-order typifications, whilst they are both arbitrary and conventional in the sense that there is no natural basis for their relations with their referents, are treated nevertheless as governed by a specific paradigm of reference which limits their meaning to a specific realm of discourse. They are not, in other words, open to a potentially infinite variety of possible meanings, but are tied to the limits imposed on them by the requirements of particular forms of discourse such as natural and physical or social sciences, philosophy or more specific epistemological disciplines and systematic practices within these forms. They are the terms of critical and reflexive theorizing and their associated methodologies, of concepts and their definitions.
A distinction related to that of Schutz is made by Garfinkel (1963: 4-6), who differentiates between indexical and objective expressions. Indexical expressions are what he terms 'contingent accomplishments of organized artful practices of everyday life' (1967: 11). They refer to phenomena which represent them in the specific uniqueness of their particular and concrete manifestations, and depend for their interpretive meanings strictly upon the contexts in which they are employed (Bar-Hillel 1954; Garfinkel and Sacks 1970: 348-50). Objective expressions, by contrast, are employed to represent the general properties of phenomena and do not depend for their meanings upon the specific contexts of the particular manifestations of those phenomena. Whereas indexical expressions are bound by the contexts in which they are employed, objective expressions are employed as if they are context-free: indeed, Garfinkel sees them as an indispensable basis for the discourses of what he terms the exact sciences. For despite the enormous utility of indexical expressions for mundane interaction, they are inappropriate for formal discourse, which depends, according to Garfinkel, on a clear distinction between and substitution of objective for indexical expressions. Yet this distinction and substitution, Garfinkel notes, remains programmatic and unsatisfied 'in every particular case and in every actual occasion in which [it] must be demonstrated. In every actual case without exception, conditions will be cited . . . such that in that particular case the terms of the demonstration can be relaxed and nevertheless the demonstration be counted an adequate one' (1967: 6). Garfinkel thus moves beyond Schutz's distinction between first-order and second-order typifications to imply a set of what he terms formal or rational properties of indexical expressions which enable them to offer a practical infinity of potential meanings whilst making possible, through their association in accounts of the phenomena to which they refer, a formal discourse which can be interpreted routinely as meaningful within much narrower limits. This is due, he argues, to the essential reflexivity of language. In accounting
307
CORE SOCIOLOGICAL DICHOTOMIES
for a phenomenon in a language, the phenomenon is in effect constituted as meaningful within the interpretive limits represented by the conventions of usage of that language. It is the language of discourse for any community that creates the common sense of that community (Filmer et al. 1972).
The relation between the semiotic and social character of image/text relations is taken still further by Charles Sanders Pierce (1867-1914) who argues the necessity for categorizing signs on the grounds that 'a sign ... is an image of its object and, more strictly speaking, can only be an idea' (Buchler 1955: 105). Pierce identifies three principal forms of sign - icon, index and symbol - into one or other of which all signs can be classified. He differentiates them as follows: 'an icon is a sign which would possess the character which renders it significant, even though its object had no existence ... an index is a sign which would, at once, lose the character which makes it a sign if its object were removed ... a symbol is a sign which would lose the character which renders it a sign if there were no interpretanť (1955: 104). It is on Pierce's concept of index that Garfinkel draws in identifying indexical expressions, as he draws on the concept of icon in identifying objective expressions. Like objective expressions, and Schutzian second-order typifications, iconic signs are abstract categories within which the class of phenomena to which they refer can be collected in terms of their common general characteristics, whatever the specific social, cultural and historical contexts in which they occur. But Pierce's concept of the symbol goes beyond these categories of Garfinkel and Schutz, and returns us to the concept of image and its relation to text. Pierce develops the complex semiotic character of the symbol in terms of its emergent relations with other types of sign, noting that its
representative character consists precisely in its being a rule that will determine its interpretant... A symbol is a law, or regularity of the indefinite future . . . it must denote an individual, and must signify a character . . . The symbol is connected with its object by virtue of the idea of the symbol-using mind, without which no such connection would exist . . . Symbols grow. They come into being by development out of other signs, particularly from icons, or from mixed signs partaking of the nature of icons and symbols ... it is only out of symbols that a new symbol can grow ... In use and in experience, its meaning grows . . . The symbol may, with Emerson's sphynx, say to man, 'Of thine eye I am eyebeam.' (1955: 112-15)
In effect, thus, the symbol is for Pierce both the originary category of sign (the eyebeam of the eye), since it is dependent on the ontological idea of the 'symbol-using mind', and also the highest order of the sign, since it is an emergent development from other signs (see Eco 1981: 175-99). In this it parallels the emergent synthesis of the dialectical interrelation between symbol and society identified by Durkheim. Similarly, in its ability both to synthesize and to authorize the meanings of other signs, it carries what
308
IMAGE/TEXT
Garfinkel terms the essential reflexivity of language. For Pierce, 'all words, sentences, books, and other conventional signs are symbols . . . The word itself has no existence although it has a real being, consisting in the fact that existents will conform to it', and they will do so because they exist 'in the possibly imaginary universe to which the symbol refers' (1955:102,112). This is the sense in which the symbol not only is an image of its object but, in being so strictly speaking, as Pierce puts it, is no more than an idea which requires realization by being transformed into a condition ('real being') in which the phenomena to which it refers will conform to that referentiality. What it requires, in effect, is to be con-texted, to be constituted in and as a text. Hence a symbol can be both a word (or other conventional sign), that is, a single unit in a text (which is also a text in itself), and also an assembly of words (or other conventional signs) that constitute a much more elaborate text, such as a sentence or a book. For Saussure, the process of contexting requires limiting what he terms la parole, the promise of potential meanings held by the terms of la langue, at the point at which it becomes langage, language in use. The potential meanings are part of a syntagmatic linguistic system which is quite arbitrarily cross-cultural and trans-historical in character. Signs in a syntagmatic system can be likened to both Durkheim's and Pierce's senses of symbol precisely because they have the potential to mean, and they will mean. But they need to be brought to the condition in which that potential can be realized in and as text. For Saussure, context is a paradigmatic system which orders assemblies of signs from potentially infinite varieties of possible meanings into the 'real being' of actual interpretive meanings fixed by convention. This locates signs reflexively, as texts, in historically and culturally specific social contexts (Barthes 1967b; Guiraud 1975). The contexts, thus, both imply the paradigmatic principle(s) which make texts meaningful and are guaranteed constitutively by the texts to the extent to which they are successful in doing so. This demonstrates, once again, the essential reflexivity of language operating through the structure of image/text relations: the text both realizes and constitutes the paradigmatic image.
In Airman's sense, a text constitutes a paradigmatic condition for itself by establishing limits to the meaning of the terms associated within it through that very structure of association itself. Hence, texts are articulated into two levels which Altman (1981: 40) terms functions and signs. The signs of a particular text can have meanings which are prior to their inclusion in that specific text. They may be more or less closely related to the meaning(s) generated by or associated with their meanings in other texts which have already been constructed, or are being created simultaneously, or may be produced in future. But the functions of the text limit the meanings of its signs to the limits which the text reflexively and simultaneously generates for them and imposes upon them. This limiting process is the process of textual interpretation. It does not imply that there is one sole meaning to a text, and hence to the functions attributed to its constitutive signs; but it
309
CORE SOCIOLOGICAL DICHOTOMIES
does indicate the manner in which an interpretation makes explicit the text's paradigmatic condition. This is, in effect, both a translation of image into text and, thereby, a concretization of the image, from its private absence from the social world in subjective consciousness, into the social and cultural presence of its public meaningfulness (Eco 1989).
It is important to note, however, that this process of self-constitution of the text at the functional level is only one of three sources of meaning for the text - what Altman (1981: 45) terms its intratextual meaning. Another source of meaning is that which derives from the words or other conventional signs of the primary language of the text - its linguistic meaning. Finally, all texts derive from their intratextual meaning an intertextual meaning which is a crucial feature of their paradigmatic condition. Intertextual meanings parallel, at a higher level of generalizability, the linguistic meanings of texts. Linguistic meanings are effectively denotative, whereas intertextual meanings are connotative, and relate the texts to the most fully elaborated paradigmatic resources for meaning in the cultures in which they are constructed and interpreted. These developed paradigms constitute what can be termed the cosmology of such cultures. As such, they are the most fundamental socio-cultural resources for collective and hence shareable images.
Barthes (1972) endorses this point in his reconceptualization of myth as a form of speech, which constructs itself as a higher-order semiological system. He terms this higher-order system metalanguage, proposing that it is a second language in the terms of which it is possible to speak about a linguistic system of a lower order. This lower-order system is composed of language objects on which myth draws to build its own system. Myth, thus, is in a reflexive relation to its lower-order system of language objects: it both draws on and speaks of them. Metalanguage enables semiologists to treat writing and pictures in the same way because both are conventional signs and hence have the same signifying function in relation to myth, which is to constitute language objects. Mythic speech, for Barthes, is speech which draws on semiotically full signifier/signified relations. The signs constituted by these semiotically full relations are already related to the paradigmatic structures of meaning which constitute cultural cosmologies: it is the contents of these structures which fill them and which give them their mythic status, no matter in which particular epistemological discourse (of art, science, religion, for example) they manifest themselves. And it is these semiotically full relations which represent the fundamental relation between image and text, since the sources of images are the cosmologies which fill the signifier/signified relations with the potential for meanings which can be interpreted within the paradigmatic limits that constitute them and make them recognizable (and interpretable) as texts. Effectively, cosmologies both provide and provide for paradigms in the sense that they are both the resource for connotative meanings which can exist in language objects prior to their constitutive incorporation into texts, and the site for the texts whose intratextual meanings enable connotations to be further elaborated.
310
IMAGE/TEXT
Barthes's argument that myth is a form of speech echoes that of Lévi-Strauss, for whom the function of signs is to 'transcend the contrast between the tangible and the intelligible . . . [through expressing] the one by means of the other' (1969: 14). If image and text are considered, respectively, as instances of the intelligible and the tangible, then the parallel becomes clear. By giving signification to the intelligible image, the sign makes it tangible and hence textual. This concept of the function of the sign is crucial to Lévi-Strauss's concept of myths, which operate autonomously in the human mind, without the awareness of the human subject, to the extent that he suggests analysing them 'as if the thinking process were taking place in the myths, in their reflection upon themselves and their interrelation . . . myths themselves are based on secondary codes (the primary codes being those that provide the substance of language)'. The structural analysis of myth implies 'a tertiary code, which is intended to ensure the reciprocal translatability of several myths ... as it were, the myth of mythology . . . each myth taken separately exists as the limited application of a pattern, which is gradually revealed by the relations of reciprocal intelligibility discerned between several myths' (1969: 12-13). From a sociological perspective, the genesis of this concept of myth as tertiary code serves a purpose comparable to that of Barthes's concept of myth as metalanguage: both provide for the conceptualization of culture as generating reflexively, through its constitutive cosmologies, the paradigms which provide for the connotative interpretability of texts (Barthes 1975).
CONCLUSION: IDEOLOGY AND IMAGE/TEXT RELATIONS
One prolific source of images in contemporary culture is the media of broadcast communication - in particular, television - whose discourses provide us with interesting sociological examples of how paradigms provide for the connotative interpretability of texts. In a pluralist political culture, the media carry messages which represent the views of different and divergent interests on the possession of power and the distribution of wealth. These views, whilst divergently different in and of themselves, are views nevertheless of what are treated as common phenomena, such as images of historical events, political and economic institutions and social activities. These phenomena, that is, are represented initially (as news, for example) as if they have occurred in the same way for all who are interested in them; but they are open to be interpreted and explained in quite different ways according to the differences between the interests of those who seek to establish their meaning. It is not actually the case that they are even represented initially as if they have occurred in the same way; rather, it is that the connotations of the conventional signs in which they have been represented have not been addressed in terms of the particular types of
311
CORE SOCIOLOGICAL DICHOTOMIES
paradigmatic resources from which the signs and their connotations have been drawn. Hall et al. (1980) formulates the processes of interpretive representation and 'reading' of the text of messages broadcast by the media as a disjunctive set of processes of systematically distorted communications (see Habermas 1970) which he terms non-symmetrical encoding/decoding. Messages are encoded according to the distortions of one system of interests and decoded according to those of another.
The systems of shared beliefs and values which underpin common interests on the use of power and the distribution of wealth in any political society are its ideologies (see chapter 'Theory/Practice'). They offer coherent and persuasive accounts of the states of the social world, of the causes, conditions and directions of societies - often as important features of their cultures. Eco notes that
the human labour of transforming states of the world . . . cannot be performed without organizing such states of the world into semantic systems. In order to be transformed, the states of the world must be named and structurally arranged. As soon as they are named, that system of sign systems which is called 'culture' (which also organizes the way in which the material forces are thought of and discussed) may assume a degree of extra-referential independence that a theory of codes must respect and analyse in all its autonomy. (1977: 297)
The process of encoding of broadcast media images is, in effect, one form of organizing interpretively a state of the world into a semantic system; the process of decoding is that of interpretively reorganizing that semantic system to generate a different organization of the state of the world. As Eco points out, the interpretive organization of semantic systems is a fundamentally social activity: 'the labour of sign production releases social forces and itself represents a social force. It can produce both ideologies and criticism of ideologies. Thus semiotics ... is also a form of social criticism, and therefore one among the many forms of social practice' (1977: 298).
Hall et al. (1980: 136-8) is thus able to suggest three hypothetical positions in terms of which the messages of a televisual discourse which has been encoded in the dominant ideological sign system of a political culture can be decoded in terms of a different sign system. These are, first, the dominant-hegemonic position in which the connoted meaning of the message is decoded in the terms of the sign system in which it has been encoded. This is the only sustainable sociological formulation of corre-spondential communication, in which the intentionally connoted meaning of the communicator is understood and accepted by the recipient of the message. The recipient is thus operating within the dominant code and participating in a process of ideological reproduction at a global level. The second position is identified as that of the negotiated code, which modifies the dominant hegemonic code of the intended connotations of the message
312
IMAGE/TEXT
by adapting it to the recipient's specific local knowledge and experience. This means that the negotiated position is one of particular exceptions and contradictions to the universality of the dominant code. Because of such contradictions, Hall notes that it is with the deployment of the negotiated code that so-called 'failures' of communication occur from the point of view of professional broadcasters utilizing the dominant code, and 'misunderstandings' from the point of view of the decoders. Thirdly, recipients can decode messages in terms of an oppositional code at a global level by replacing the connotative system of the dominant code with a completely different and divergent system of connotations. It is at this point, Hall notes, that 'the "politics of signification" - the struggle in discourse - is joined' (1980: 178). It is at this point, too, that the labour of sign production becomes a social force by generating ideological and social criticism. Just as the dominant code has attempted to reinforce the existing socio-cultural order by connoting its hegemony in broadcast messages, the image of an alternative state of the social world has been connoted as a realizable possibility by being encoded as the text of a different sign system. As social forces, and as practices of social criticism and the criticism of ideologies, thus, the dichotomous semiotic and linguistic structures of relations between image and text are central to contemporary critical sociological inquiry.
KEY CONCEPTS
IMAGE Although an image is a likeness the word has come to suggest a gap between the real and the created (the image). The image is thus more like the imagined, however, as sociologists know, images and imaginings can be real in their consequences so we have to seek to understand the reality both within and behind any image.
TEXT Text is a way of talking about any reality as a series of potential messages. Psychologies, personalities, identities, books, films, social events can all comprise texts that can be read in a variety of different ways. It is critical for the social theorist to understand the rules and interests contained within any particular reading and to justify any reading of their own.
Realism An aesthetic which argues that the role of art is to represent reality. It is developed most fully during the nineteenth century, alongside materialist philosophy and  scientific knowledge, and in  opposition to
313
CORE SOCIOLOGICAL DICHOTOMIES
philosophical idealism. Amongst its principal consequences is a correspondential theory of linguistic and semiotic representation, according to which the work of art is a reflection of the reality of experience. The structure of social reality is seen, in turn, as determining the work of art.
Hermeneutics The term was imported from theology into modern philosophy by Wilhelm Dilthey (1833 - 191 1) to characterize the interpretive investigation of essentially intentional human acts of linguistic understanding. It is also used by Heidegger to denote the phenomenology of existence and understanding. Ricouer argues that contemporary hermeneutics is concerned with two principal forms of textual exegesis: by recovery, in which the text is analysed interpretively to reveal a meaning hidden within it; and by suspicion, in which the text is treated as concealing a set of cultural determinations which must be exposed.
Ontology That part of metaphysics which is concerned with inquiry into the conditions and character of (human) being itself. It has become important for modern philosophy as a result of the phenomenological critique of positivism. Positivism sought to end metaphysical speculation altogether, and subsumed ontology into epistemology, from which it had been traditionally differentiated, thereby proposing to establish being as (human) nature. Phenomenology reestablishes ontology at the centre of philosophical inquiry because of its focus on the reflexive character of consciousness.
Sociologism A form of reductionist explanation of all human social phenomena in sociological terms. It is most closely associated with Durkheim's sociology, though as Tiryakian (1962) notes, he neither defines nor uses the term. As a type of sociological explanation it depends upon the differentiation of social from all other types of human phenomena, and it insists upon their irreducibility to explanation in terms of other epistemological discourses, such as anthropology, history or psychology.
Semiotics/semiology The systematic study of signs and their denotative and connotative properties. The term was used by Pierce at the end of the nineteenth century to represent a general theory of the logical function of signs. He proposed that semiotics was effectively another name for logic in its general sense and thus a science of what he termed abstractive observation. Shortly after this, Saussure conceived the term semiology to represent the study of the social character and functions of signs.
314