Reading guide to selections from: Barthes, R (1977) Image-Music-Text, London:Fontana Press

The Death of the Author

First, the piece offers an uncompromising argument for the 'death' (or redundancy, to use an equally frightening metaphor) of the concept of the author. It begins by noticing the disappearance of the narrator in modern writing . Balzac's Sarrassine is the example here, later to be the subject of a much larger piece by Barthes (1975) (the dates refer to different English editions of Barthes' work, and are not reliable as a guide to the actual sequence of the writing). Perhaps something like Martin Amis's London Fields, with its switches between different narrators, might be more appropriate an example for the modern reader? The effects of surrealist or Brechtian experimentation are also cited, as steps on the way, so to speak. There is an insistence that even autobiography is not about real life coded in writing but the other way around -- e.g. Proust reconceptualised his life after or during writing in order to make it a 'work for which his own book is the model' (Barthes 1977:144). As Sturrock (1979) reminds us, Barthes wrote his own autobiography in the third person. 

Secondly, as structuralist linguistics tells us, texts do not express the subjectivity of their authors -- they are better thought of as 'fields without origin', 'multi-dimensional spaces', 'tissues of quotations', 'never original' (Barthes 1977: 146). This point applies to all kinds of mundane feelings of 'authorship' as well as actual novel-writing. The inner self that we experience as the 'real us' so to speak, is 'only a ready-formed dictionary' -- so life imitates books. 

There are no fixed meanings or privileged ones: 'writing ceaselessly posits meaning ceaselessly to evaporate it' (147). We should see the act of writing as 'performative'. As a result, conventional literary criticism, designed to uncover the 'real' meanings of novels, expressed by 'real' authors, is also abolished and that's good, because, in its arrogance, such criticism used to ignore the reader and also selectively overlook the 'phatic' bits of texts (designed to involve the reader). 

The reader is the missing term in conventional criticism -- the multiplicity of the text is focused in the reader, not the author, the unity of the text is in its destination. '[T]he birth of the reader must be at the cost of the death of the author'. Yet, to raise a point which we will discuss below, 'the reader' is also an abstraction 'without history, biography, psychology' (148). Really, of course, it could not be otherwise for Barthes: it would be inconsistent to abolish the (actual) author while retaining the (actual) reader. 

From Work to Text

This piece begins with an interesting comment about the effects of a move towards interdisciplinarity 'when the solidarity of the old disciplines breaks down -- perhaps even violently' (Barthes 1977: 155). We now have an escalating pattern of change, where the classic 'breaks' (attempts to re-found marxism, Lacanian freudianism, and structuralism) have relativised our knowledge of the world, and changed our notions of the relations between writers, readers and observers. There will be further changes in the basis of knowledge, a new 'epistemological slide' rather than a break (155). (Althusser had claimed to have found an 'epistemological break' between humanist and structuralist marxism in Marx himself, a break that inaugurates a new 'scientific' marxism). 

For Barthes, the new (literary and cultural) analysis will not be a new, tightly ordered discipline, but should be seen as necessarily speculative, employing 'not argumentations but enunciations, "touches", approaches that consent to remain metaphorical' (156), rather than an attempt to read off meaning from a metalanguage. Methodologically, we are told (164): 

'the discourse on the Text should itself be nothing other than text, research, textual activity, since the Text...leaves no language safe outside, nor any subject of the enunciation in position as judge, master, confessor, decoder. The theory of the Text can coincide only with a practice of writing.' 
We see here one of the ways in which Barthes is shifting to a radical kind of textuality, then, one which 'goes all the way down' as Norris (1992) puts it, and one where Barthes, like Lyotard heads into a denial of any other criteria by which to judge events or accounts of them. We get to the heart of Barthes' argument when he tries to distinguish 'texts' and 'works'. The distinction is not a matter of location in time or of value or quality. Works might have different qualities among themselves, but this is not the proper topic for a critic: -- 'there is no difference between "cultured" reading and casual reading in trains' (162). By implication, for Norris, there is also no difference between such works and those in history or economics either. 

Works are designed to be consumed with plaisir (roughly, a rather conformist pleasure delivered by a work, gained from following the narrative to its delivery point and responding as intended). Works are 'filiated', closely connected with social practices like those in the world of commercial writing, including literary criticism and notions of authorship, ownership, copyright and the law. (This point is made in Foucault's own announcements of the 'death of the author' --collected in Bocock and Thompson 1992). 

By contrast, texts are distinguished by their 'methodological fields' rather than by anything substantially or concretely different about them or their contents. They are performances, 'limit works', existing at the limits of 'enunciation, rationality, readability etc.' (157). Texts are radically symbolic, 'off-centre, without closure (159), playful, offering jouissance (an ecstatic pleasure in language that escapes the devices of the narrative and rejoices in the experience itself, a kind of literary orgasmic release -- see Heath's discussion in the introduction to Barthes 1977). Texts help us glimpse a 'social Utopia...[a]...transparence of linguistic relations if not social ones', a 'space where no language has a hold over any other' (164). Texts operate via 'serial movements of disconnections, overlappings, variations' (158). 

Texts are networks (rather than discrete entities or 'organisms' with a history and parentage like works), opening out to readings well outside the author's intentions. Such openness blurs the conventional differences between reading and writing (162). The reader and the text both play with meaning, rather as a musician plays with a score, both to reproduce it and to embellish, to perform. There is here a clear preference for texts, of course, and a way of denying any claims to sufficiency advanced by any mere works 

Change the Object Itself

The old critical project, as in Barthes' earlier classic Mythologies (Barthes 1973) has to be altered. That earlier work followed an 'inversion model' (common in marxism, but subsequently attacked by the Althusserians -- see file), which saw connoted ideological meanings as a base for the denoted literal meanings of cultural phenomena. In this way the meanings of advertisements, performances, cultural activities of various kinds naturalised capitalist ideology. 

As an aside, it is worth looking at some of the pieces in Mythologies, perhaps. My personal favourite concerns a very brief analysis of what became known as the 'structure of apology'. At the most specific level, the piece, entitled Operation Margarine, features an analysis of an advertising campaign for margarine which cleverly acknowledged and then incorporated the consumers' perceptions of margarine as an inferior product (modern readers might think of substituting junk food, say, for margarine in order to locate themselves in the politics of the piece). 

Apparently, the advertisement began with a 'cry of indignation against margarine: "A mousse? Made with margarine? Unthinkable!"' (1973: 42). A narrative then developed which revealed these perceptions as misguided and ill-informed: 'And then one's eyes are opened, one's conscience becomes more pliable...The moral at the end is well known: "Here you are, rid of a prejudice which cost you dearly!"' (42). Barthes sees this sort of structure operating in the wider society: 'It is in the same way that the Established Order relieves you of your progressive prejudices' (42). He finds the principles at work with discussing the Army or the Church, for example (41) (and implicates some popular novels in the process): 


'...[on] the Church: speak with burning zeal about its self-righteousness, the narrow-mindedness of its bigots, indicate that all of this can be murderous, hide none of the weaknesses of the faith. And then, in extremis, hint that the letter of the law, however unattractive, is a way to salvation for its very victims, and so justify moral austerity by the saintliness of those whom it crushes (The Living Room, by Graham Greene).' 
This sort of analysis had made Barthes very influential, and perhaps it is easy to see why from the example above: a mundane advertisement is made to yield some concealed truth about capitalism following a skilled and sceptical reading which refuses to simply follow the account of the world on offer, but which imposes its own. Now, to return to Change the Object Itself, Barthes feels it is time to move beyond this sort of analysis of mythology as a mystified or an upside-down world, as inversion. This analysis was drawn from themes in the young Marx, but we can now progress to those in the mature Marx (another specific reference to Althusser's project here, Barthes 1977: 169). 

Mythology still works in the same way, but we now have a new science of reading it. Also: 'any student can and does denounce the bourgeois or petit bourgeois character of such and such a form' (166). Denunciatory discourse and demystification have been routinised, have become a mere 'stock of phrases', orthodox, even mythological themselves. As a result, properly academic and critical analysis must go further and 'shake the sign' itself, just as French psychology has moved on: that began by listing the symbolic contents of dreams and so on, and 're-inverting' them, only to find that being done these days by mere dabblers, the 'psychological vulgate' (167). 

The task now is not to reveal latent meanings but to 'fissure' meaning and its representation, not to destroy myths ('mythoclasm'), but to splinter the smooth connections between signs ('semioclasm'), not to critique just French society but the whole of Western civilisation and its unifying 'regime of meaning' (167). This is a project to dissolve any 'works' back into 'textuality', in other words. 

Apart from being made possible by the new methodological work available, this shift is necessary because the signs of advertisements no longer point to products nor to political ideologies as simply as they did. The whole world is already playing with signs: 'endlessly deferring their foundations, transforming signifieds into new signifiers, infinitely citing one another' (167--8). The issue now is not one of 'critical decipherment' but of estimating the 'levels of reification of various languages, their 'phraseological density' . 

We have an interest, then, in the extent to which these different 'languages' can appear as fixed, immutable, natural and compelling, as 'works' making claims to be sufficient, to use the terms we developed above. 

This is a shift from a more obviously 'denunciatory' stance, with its problems of separating out the error of the myth from the truth of its analysis. In a way, Barthes here is anticipating the failure of the Althusserian project, perhaps, which was the last great attempt to clarify the basis for marxism's claims to be able to offer a 'science' to help us identify 'ideology' or 'myth' (one of the last great metanarratives in Lyotard's terms). 

Myth is still universal in our societies, affecting 'inner speech, newspaper articles ... political sermons, [running] from the novel to the advertising image (i.e. all the imaginary)' (169). We need new concepts to grasp it, not the old ones of sign, signifier, signified, connotation and denotation, but 'citation, reference, stereotype'. We need to offer an 'antidote to myth', and its reifications, languages which are 'airy, light, spaced, open, uncentred, noble and free' (168), a 'new semiology'. 


Barthes, R. (1973) Mythologies, London: Paladin.

Barthes, R. (1975) S/Z, London: Jonathan Cape

Bocock, R. and Thompson, K (eds) (1902) Social and Cultural Forms of Modernity, Cambridge: Polity Press.

Norris, C. (1992) Uncritical Theory: Postmodernism, intellectuals and the Gulf War, London: Lawrence and Wishart.

Sturrock, J. (ed) (1979) Structuralism and Since: From Levi-Strauss to Derrida, Oxford: Oxford University Press.