One: Approaches to Discourse Interpretation

One The Pretext of Interpretation


University of Vienna, Vienna, Austria


It is a common assumption, taken as self-evident in much work on critical discourse analysis, that significance can be assigned to texts by means of an analysis of their linguistic features. But texts are only naturally produced and processed in conjunction with particular contexts and pretexts, so to assign meaning to them in isolation is pragmatically invalid, based on a conceptual confusion between analysis and interpretation, between signification and significance. Meanings are not discovered in texts but derived from them and are therefore always conditional and indeterminate, with pretext the key regulating factor.

I will begin by stating the obvious and then go on to consider implications which, it would seem, are not so obvious. When people want to relate to others for one reason or another-to communicate with them, persuade them, co-operate with them, inform or instruct them, impress or oppress them or whatever-then they make use of language as a convenient means for doing so. They draw on the linguistic resources at their disposal to produce texts of one kind or another-articles, reports, manifestoes, public notices, cooking recipes, letters and lectures-like this one. I am producing text as I speak so as to relate to you. Texts are then a convenient means for mediating between people. But how does this mediation work? Texts are composed of linguistic signs in combination, so what do these signs signify, what significance can we attach to them?

As I say in the abstract to this talk, it is a common assumption, taken as self-evident in much work on critical discourse analysis, that significance can be assigned to texts by means of an analysis of their linguistic features. I want to argue against this assumption and to suggest that it is not only misconceived, but that it distracts attention from what I think are the essential issues about the use of language in communication. No amount of analysis, I shall argue, no matter how precisely it is carried out, can reveal the significance of texts, and indeed the more precise the analysis, the less revealing it will tend to be.

Texts are produced, as visual marks in writing or sounds in speech, by a first person writer or speaker, P1 for short, and received and processed by second person readers or listeners, P2 for short. I am a P1 producing text now, of course, as I speak to you P2s, and you are, I hope, processing it. Critical discourse analysis claims that the linguistic features of a text are indicative of the ideological position of the P1, but that these indications are not apparent to ordinary second person readers/listeners, the P2, who lack the analytic skill to identify them and who therefore fail to realize their significance. Thus the ideology is subtly insinuated into the readers’ minds. The task of CDA, indeed the cause that it embraces, is to expose such covert and subversive intentions by close textual analysis and make the reader aware and wary of the persuasive power of texts and the deception that is being practised on them. The cause is, of course, a worthy one. How language is used to persuasive effect by those in power, whether for political or commercial purposes, is something that warrants serious and critical investigation. My argument, though, is that the procedures of CDA, or at least those that are widely followed, are unhelpful to this investigation because they are based on a misconception about the nature of text and a confusion between analysis and interpretation.

What then is a text? It is a linguistic object, a manifestation of language that can be anything from a one word public notice to a scholarly monograph. Texts can be analysed in isolation, and their lexical and grammatical features described, nowadays with the aid of computers to a high degree of precision. But though they can be analysed in isolation by linguists, they are never produced or received in isolation by language users. When we produce or process a text, in speech or writing, we are prompted by some reason or other: P1s do not just produce texts unprompted out of the blue and P2s do not just pick up texts at random. There is, in short, always a pretextual purpose.

In fact, users do not really experience texts, any more than they experience sentences. Writers do not write texts and readers do not read them, they write and read notices, letters, articles, reports, monographs, poems, novels and so on. Texts are simply the linguistic trace of a pragmatic process whereby the writer, the producer, the first person or P1 uses language indexically to make a contextual connection for some communicative purpose or other. It is this pragmatic process of meaning realization that I refer to as discourse. Language users enact discourses by means of texts, and in this two other factors are crucially implicated: context, the extralinguistic reality that linguistic features point to, and pretext, the intended purpose of P1 that motivates the communication in the first place.

All this sounds, and indeed is, somewhat abstract, so let me give you an example. Here is a piece of text-an old text, but not any old text. It is one of the most cited and celebrated in history: the opening of the

American Declaration of Independence.

We hold these truths to be self-evident, that all men are created equal, that they are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness.

If we consider the phrase “all men” we can assign it a semantic meaning without difficulty since we know the denotation of each constituent word. But what the phrase is intended to refer to in this text is a very different matter-a pragmatic and not a semantic matter. Is the reference meant to override the denotation of “men” and include women as well? Given the context of the time, probably not. But what of the denotation of “all”? All men. All men. Thomas Jefferson and the other founding fathers who drafted this document were, of course, slave owners, and they certainly did not hold it as a self-evident truth that slaves were divinely endowed with an unalienable right to life, liberty and the pursuit of happiness. And there were other men, too, who are referentially excluded-the indigenous inhabitants of the continent, who, later in the document are referred to “the inhabitants of our frontiers, the merciless Indian Savages”. The self-evident truth that all men are created equal does not apply to them: no unalienable rights for aliens such as these. So it is clear that what the phrase “all men are created equal” means semantically is not at all the same as what the authors of this text meant by the phrase. What it is intended to mean is not all men, but all men of a certain kind- men like Jefferson in fact. The first person “we” and the third person “all men” are in effect co-referential.

So the historical context here determines the scope of reference intended by the writers of this text. But we can relate this text to a different context and so change the referential scope to include all human beings, slaves, indigenous people, men as well as women. And this indeed is what happens when these words are cited these days, and used to express the principle of universal human rights. The phrase “all men” is now taken to mean all human beings, without exception, which is not at all what the original writers of the text intended by it.

But it is not only that the text is now related to a different context. It also serves a different purpose, and here pretext comes into play. This text is part of a proclamation, a declaration of independence and is simultaneously directed at two P2 recipients: like-minded fellow American colonists on the one hand, and the British colonizers on the other. So the declaration is at one and the same time designed to the expression of common aspiration and an act of defiance directed at the colonial power. And if it had the desired effect on the readers for whom it was intended that was really all that mattered. In other words, the writing of the text was motivated by a pretextual purpose to bring about a certain effect on certain readers who could be counted on to recognize and ratify this purpose. They would not subject it to close reading, certainly not to linguistic analysis. For them it would not be a text as such at all, but a declaration, an act of defiance, a call to arms in a political cause and they would regulate their attention to the text accordingly. And as with context, we can of course, relate this text to a different pretext, as when it is cited, as it frequently is, not as a revolutionary incitement to rebellion but as a declaration of the democratic principle of universal human rights. And this, we should note is in effect directly contrary to the pretextual purpose of the signatories to the Declaration of Independence which was essentially to invoke the principle as a convenient tactic to assert their own particular rights.

So what the writers of this text do is to make expedient pragmatic use of the semantics of English to suit themselves. They are not the only ones. This is 1776. Fast forward 250 years and we still find the same thing: language routinely pressed into the service of a political pretext. Here is Chomsky:

When Western states and intellectuals use the term ‘international community’, they are referring to themselves.... Those who do not support the actions of wealth and power are not part of the ‘global community’, just as ‘terrorism’ conventionally means ‘terrorism directed against us and our friends’. (Chomsky 2001)

So like the term “all men” the term “international” is used pragmatically to mean what it suits its users to mean, in defiance of its semantic denotation.

This, to change the context rather abruptly from the real world to the realm of fantasy, is also what Humpty Dumpty does-a character in Lewis Carroll’s Alice Through the Looking Glass. He uses the word “glory” and tells Alice that it means “a nice knock-down argument”. Alice objects that this is not what the word means, and here is his reply:

‘When I use a word,’ Humpty Dumpty said in rather a scornful tone, ‘it means just what I choose it to mean-neither more nor less.’

‘The question is,’ said Alice, ‘whether you CAN make words mean so many different things’

‘The question is’, said Humpty Dumpty, ‘which is to be master-that’s


Obviously we cannot master words to the extent that Humpty Dumpty claims we can. Using language as it suits us cannot extend to assigning to words any arbitrary meaning we choose. There has to be some convention in invention, some semantic common ground; otherwise texts could not mediate between P1 and P2 in the communication process at all. Without such consensus about encoded meaning, there can be no convergence. But the question is not, as Alice has it, whether we can make words mean so many different things. Clearly we can. But to what extent and in what circumstances? How far does what words mean semantically constrain what we can mean by them? We can make words mean many different things pragmatically, but the question is what are the contextual and pretextual conditions that enable us to do so.

What I am saying, then (and it is obvious enough when you think about it), is that texts do not vary but their interpretation very definitely does. In other words, texts are fixed and determinate linguistic objects but they give rise to variable discourses. To return to the Declaration of Independence. This must be one of the most cited texts ever written. And each time it is cited, it is the same text. But the point is that though the text remains the same, its significance does not because the contexts and pretexts it relates to change depending on when it is cited, by whom, for what purpose. So although we can analyse the text linguistically in terms of its lexis and grammar, describe the meaning that is semantically encoded in its sentences, this will not tell us what discourse it is a trace of, or what discourse its readers derive from it.

And of course the discourse that the first person writer, the P1, intended to textualize may well not correspond at all closely, or even at all, with the discourse that the second person reader, the P2 derives from the text. What writers have in mind may not at all transfer to the minds of their readers. The novelist Doris Lessing has interesting things to say about this in the preface she wrote for the reprint of her novel The Golden Notebook.

Ten years after I wrote it I can get, in one week, three letters about it....One letter is entirely about the sex war, about man’s inhumanity to woman, and woman’s inhumanity to man, and the writer has produced pages and pages all about nothing else, for she-but not always a she-can’t see anything else in the book.

The second is about politics, probably from an old Red like myself, and he or she writes many pages about politics, and never mentions any other theme.

These two letters used, when the book was, as it were, young, to be the most common.

The third letter, once rare but now catching up on the others, is written by a man or woman who can see nothing in it but the theme of mental illness.

But it is the same book.

And naturally these incidents bring up again questions of what people see when they read a book, and why one person sees one pattern and nothing at all of another pattern, how odd it is to have, as author, such a clear picture of a book, that is seen so differently by its readers. (Lessing 1972: xix-xx)

Doris Lessing is of course talking about a literary text and literary texts, not being so tied to contextual constraints are particularly prone to variable interpretation-they are indeed designed to be, and this is essentially their pretextual purpose, as Lessing herself recognizes. She goes on to make what she calls “a most fundamental point”:

Which is that the book is alive and potent and fructifying and able to promote thought and discussion only when its plan and shape and intention are not understood, because that moment of seeing the shape and plan and intention is also the moment when there isn’t anything more to be got out of it.

In other words (in my words), if the meaning of a text were to be fixed, no discourse could be derived from it and it would in effect cease to function as a text.

But what Doris Lessing says applies to all texts, not just literary ones. All texts are subject to variable interpretation and they would not function as texts otherwise. They always presuppose some context and some pretext, and it is only because these are presupposed that texts exist at all, and only by reference to these factors that they can be interpreted. As I have said, they are never produced, or received in isolation, and so to isolate them as linguistic objects for analysis is necessarily to misrepresent them. So really it makes no sense to ask what a text means. It does not in itself mean anything. What we have to ask is what do writers mean by their texts and what do texts mean to their readers.

Now of course writers will always rely on readers’ shared linguistic knowledge to serve as a common semantic base and so the linguistic features of the text will give indications of writer intentions and some of these will be easy enough to identify. As I have already said, there will always be some consensus about what writers are talking about and it is this that leads to the illusion that there is meaning in the text itself, but to say that a text means such and such is really only shorthand for saying that we can understand what the writer meant by it. Some intentions are clearly signalled and easy to identify. Others however, are not. And anyway, we do not normally read a text with a view to discovering what the writer meant by it. Our main concern is what it means to us. So in reading a text, we do not subject every linguistic feature to analytic scrutiny to try to find out just what the writer might have intended by using it-we would not do much reading if we did. We focus selectively on some features and disregard others, regulating our attention according to what our own pretextual purpose is in reading the text in the first place-what we are reading it for.

This view of the nature of text and of the reading process that I am proposing here is not one that is favoured by critical discourse analysts. On the contrary, the prevailing assumption here seems to be that writer intentions can be traced in any and every linguistic feature if one only looks closely enough and that if one does not do this one is missing the essential significance of the text, what ideas, beliefs, values are really being conveyed by it. Every text is, of course, the result of a selection of words and structures available in the language code. The question is, what significance can we, or should we, attach to the selection. For critical discourse analysts, the tendency is to suppose that every selection can be charged with ideological significance. Let us consider an example. In his influential book Discourse and Social Change, Fairclough sets out to analyse a particular newspaper headline to demonstrate how it expresses ideological values that the unwary reader would fail to notice. This is the headline:

Gorbachev Rolls Back the Red Army Fairclough comments:

We might well see here a different ideological investment from other ways of signifying the same event, for example ‘The Soviet Army Reduces its Armed Forces’, or ‘The Soviet Army Gives up 5 Divisions’.

We might well see this different ideological investment, but equally we might well not-it depends on who we are. Fairclough does not actually explain what change in ideological investment is brought about by these alternatives. He notes that the original “signifies a process of a particular individual acting physically (note the metaphor) upon an entity”. So we might infer that the point being made is that in the original the event is represented as brought about by a physical action on the part of Gorbachev, whereas in the alternatives, the agency is represented as an impersonal institution, the soviet army, now promoted to subject position where its agency is supposedly given prominence. So what significance are we to read into this? Institutions themselves cannot act as agents: there has to be some person or persons to make decisions and this is disguised in the alternative headlines which represent the army as an agency in its own right, an independent force over which Gorbachev has no control. Not only has he lost his agency status, and is no longer in control of his army, but he has disappeared from the scene altogether. He has become superfluous. So as far as “ideological investment” is concerned, we might well see these two ways of signifying the same event as highly significant in that they represent radically different power relations between Gorbachev and the soviet army. We might well see this. But equally we might not. I have no idea whether Fairclough himself would agree with this interpretation. And we have so far only considered one or two linguistic features. Fairclough himself points to other features we need to take note of:

‘Gorbachev’ is topic or theme of the clause, as the first part of a clause usually is: the article is about him and his doings. On the other hand, if the clause were made into the passive, that would make ‘the Red Army’ the theme: ‘The Red Army is Rolled Back (by Gorbachev)’. Another possibility offered by the passive is the deletion of the (bracketed) agent, because the agent is unknown, already known, judged irrelevant, or perhaps in order to leave agency and hence responsibility vague. (Fairclough 1992: 75-76)

The theme of a clause is, of course, a formal feature and is not necessarily to be interpreted as having the function of topic, so we cannot infer from the wording of the headline that the following article is about Gorbachev and his doings. The passivization and especially the omission of agency might perhaps be interpreted as a way of leaving responsibility vague and so changing “ideological investment”. Perhaps, but, equally, perhaps not.

And what of other linguistic features? What change in “ideological investment” would be signalled, for example, by a change of terms of reference, from the Soviet Army to Soviet Forces, for example, or Red Army to Soviet Army-are the terms marked with different connotations, with the Soviet Army a more neutral or objective term, the Red Army suggestive of menace, perhaps, or heroism? Or what if Gorbachev is replaced by Mikhail Gorbachev, or Soviet President, or Soviet Leader? Would this indicate that his action is more or less individual or institutional? And what of the verb phrase? What if we were to replace reduce with cut, rolls back with cuts back, or reduce with make reductions? So we can think of all kinds of other ways of signifying this event:

Mikhail Gorbachev Rolls Back Soviet Forces Mikhail Gorbachev Cuts Back Soviet Army Soviet Army Cut Back by Gorbachev Soviet Forces Reduced by Gorbachev Soviets Reduce Red Army Armed Forces Reduced by Soviets Reductions Made in Red Army Gorbachev Makes Reductions in Soviet Forces

And so on through a whole range of permutations, to each one of which we could no doubt assign a different “ideological investment” if we were ingenious enough.

But the obvious point is that I would not normally indulge in such ingenuity. When I come across a headline, it is in a newspaper and I read it as a headline and regulate the attention I pay to it accordingly. I do not analyse it as a text in isolation to try to discover what possible ideological significance might be assigned to it. It is pointless to try, for the possibilities, as we have seen, are endless. And this is always going to be the case whenever texts are dissociated from the contexts and pretexts that they are related to in the natural pragmatic process of making meaning.

What users of language pay attention to when they interpret texts and what linguists do when they use texts as data for analysis are two very different things. Interpretation is subject to contextual and pretextual conditions and analysis is not. This is particularly clear in the case of ambiguity. Take the famous example:

Visiting aunts can be boring.

This is ambiguous, as linguists point out, because the surface sequence of forms fuses two distinct structures: To visit aunts can be boring/Aunts who visit can be boring. But it is the sentence in isolation that is semantically ambiguous. But what if this sequence were to be used pragmatically as part of a text? Let us suppose, for example that it occurs in a letter to a friend:

Aunt Anna came round to see us again and spent a lot of time telling us in tedious detail about her holiday in Brighton. Visiting aunts can be boring.

The ambiguity disappears. The recipient of the letter is not going to notice it. Linguists may notice these things, but only by assuming a nonreader role. My favourite example of this is the comment made to me by a grammarian colleague many years ago about an expression that occurred in a news item about stormy weather on the East coast of England. The text went something like this:

Severe storms hit the East coast yesterday with winds reaching hurricane force. Off the coast of Lowestoft, five people were lost in a rowing boat.

“Five people were lost in a rowing boat”-that is ambiguous, said my colleague. It could mean “a rowing boat was lost with five people in it” but it could also mean “five people were lost inside a rowing boat” (presumably five very small people). Well, yes, it could also mean this but that is not what it is likely to mean to the readers of the report, primed as they are by the context and assuming that the writer’s purpose was to provide information rather than play with words to make a joke. The point is, of course, that the semantic ambiguity that is assigned to the sentence by analysis is pragmatically over-ridden. The context and pretext do not provide the conditions for activating it.

The problem with a good deal of critical discourse analysis is that how semantic meaning is encoded in linguistic forms is confused with how these forms are pragmatically interpreted. Ambiguity would not only be noticed but assigned significance as expressing some underlying attitude, some ideological point of view. This confusion is apparent in the following statement by Michael Stubbs:

Much text analysis, especially within critical linguistics, starts with the Hallidayan assumption that all linguistic usage encodes representations of the world. It is always possible to talk about the same thing in different ways, and the systematic usage of different syntactic patterns encodes different points of view. (Stubbs 1996: 130)

What Stubbs is referring to here is Halliday’s assumption that the grammatical features of a language are what they are because they reflect the social functions that they have evolved to serve. But this is an assumption about the language code and its historical development, not about its current usage. One can accept that, historically, linguistic forms are functionally motivated in that they have semantically encoded representations of the world. But as we have seen semantic encodings do not get directly projected in actual usage. How language functions pragmatically here and now is a very different thing from how the functioning of language in the past has become semantically encoded. What the user of a particular syntactic pattern means by using it is not at all the same as what meaning a particular syntactic pattern encodes. And we do not understand what is meant by a particular syntactic pattern, or what it means to us, by just decoding it.

Critical linguists (or many of them, at any rate) do, however, seem to understand texts in this way and so to fall prey to what I have referred to elsewhere as the functional fallacy. This is the assumption that semantic signification is directly projected as pragmatic significance in language use, and that therefore what somebody really means by a text can be recovered from the text itself if one is perceptive enough to read the signs. It is further assumed that readers generally lack this perception and so need to be told what texts really mean and what texts ought to mean to them. Fortunately, critical linguists are on hand to provide an expert exegesis by analysis to put readers right.

But, as I have argued, and I hope demonstrated, texts never contain meaning. A collection of papers by the late John Sinclair, distinguished linguist and pioneer in corpus analysis, bears the title Trust the Text (Sinclair 2004), and if one is concerned, as he was, to describe the language that people actually produce in the discourse process, this is an entirely valid piece of advice. But the trust is misplaced if one is concerned with the discourse process itself, the pragmatic use of linguistic resources to make meaning. Here, the last thing you do is to trust the text to tell you what is going on. Trust the text by all means if you are concerned only with text analysis. But if you are concerned with discourse interpretation-Distrust the Text.

For discourse meaning is never discovered in text but always to some degree invented. Significance is never simply signed but always assigned and this can only be done by taking context and pretext into account. And of course, the significance that critical discourse analysts assign to texts is no exception. These analysts too have their pretextual purposes. Although they may claim that they can reveal what texts really mean, what they actually do is to provide a commentary on what certain texts mean to them, and they focus on whatever linguistic features suit their purpose. To their credit, they acknowledge their pretext, and state quite explicitly that they are in the business of discourse analysis for ideological reasons-this, for them, is what makes discourse analysis critical. As van Dijk puts it:

Critical Discourse Analysis (CDA) is a type of discourse analytical research that primarily studies the way social power abuse, dominance, and inequality are enacted, reproduced, and resisted by text and talk in the social and political context. With such dissident research, critical discourse analysts take explicit position, and thus want to understand, expose, and ultimately resist social inequality. (van Dijk 2001: 352)

This is, as I said at the beginning of this talk, a worthy cause, and one I would myself wish to endorse. But the problem with this activist agenda is that being critical in this sense necessarily invalidates the claims that are made for the analysis. For this pretext, like any pretext, is bound to result in a partial and prejudiced interpretation which has no more claim to being real or revealing or significant than any other. Like the readers of Doris Lessing’s novel, critical discourse analysts will read into the text whatever meanings suit their own pretextual views of the world. So long as we recognize this, and read the work of critical discourse analysis as what it is, namely critical discourse interpretation-simply as inventive, interesting, thought provoking interpretations, no harm is done. On the contrary their very partiality can provide us with insights into possible ways of reading meaning into texts we would not otherwise have been aware of.

The problem is that these interpretations are presented as having a special authority because they are supposedly based on expert linguistic analysis and should therefore take precedence over any other interpretations based on different pretexts. But as I have argued, expertise in linguistic analysis provides no privileged authority to determine what writers or speakers mean by texts or what they should mean to readers and listeners. Text analysis, always necessarily selective, will support whatever partial interpretation fits the pretext of the analyst. Distrust the text and distrust the text analysts if they are pretending to be authorities on interpretation.

‘The question is,’ said Alice, ‘whether you CAN make words mean so many different things.’

Well, Alice, Yes we can. And that phrase itself provides some proof of it-‘Yes we can’, used by Barack Obama as text in association with a certain context, and a certain pretext can be assigned particular significance. And as with that other famous and much repeated phrase ‘All men are born equal’ the significance will vary. ‘Yes we can’ also served as a clarion call for many at the time of the US Presidential election-although for others of different pretextual persuasion, it had a very different effect. And what the phrase means to people now, even those who found it inspiring at the time, is again likely to be different. Again it is obvious that the significance is not in the text itself but a function of how it relates to variable contextual and pretextual conditions. That is the point. Can we make words mean many different things? Yes, we can, but what is of interest is not that we do it, but how we do it. And in investigating how we do it we can be led to understand how communication really works, how it actually depends on meanings not being semantically fixed, how the creative process of making meaning pragmatically crucially depends on the essential indeterminacy of language.

This, I think, is what Doris Lessing is getting at in making what she calls “the most fundamental point”. If an understanding of a text is taken as fixed, “there isn’t anything more to be got out of it”. She is referring to a novel: a literary text. And with literary texts understanding is particularly elusive since such texts do not key into context and pretext in conventional ways. All the more reason, one might suppose, to allow for variable interpretation. But how much variable interpretation do teachers of literature actually allow? There is still, I think, the tendency for students of literature to be told what texts really mean on the authority of literary critics, and discouraged from exploring what these texts might mean to them. Literary critics, like critical discourse analysts, claim a privileged authority to pronounce on significance, presenting what they read into a text as what the text really means.

With literary texts, students are led to believe that they cannot understand them properly unaided. With other, conventional, texts they are also often led to believe that meaning is in the text itself, there to be discovered, but in this case discovery is thought to depend mainly on their degree of linguistic rather than literary competence. Take the kind of comprehension exercise, for example where students are presented with a text in isolation and required to answer questions about what it means, without being cued into any context or pretext that would normally accompany any text, and without any pretextual purpose of their own. As I have said, nobody normally is called upon just to read a text in isolation, out of the blue. There is always a pretext and this pretext, as I have suggested, naturally regulates how much attention you pay to the text-you do not process all of it, every linguistic detail. But this is what these students are being asked to do so as to find meaning in a text that is not actually there to be found. No wonder they find it difficult. They can no doubt do some decoding of its linguistic features, but, as we have seen this is a process of semantic analysis and not pragmatic interpretation. So the misunderstanding about the nature of text and the pragmatics of interpretation that I have been discussing is not confined to critical discourse analysis but seems to be prevalent in literary studies and language education generally.

Distrust the text. Let me be clear that I do not mean by this that we should disregard the text. We obviously cannot do that because in many cases this is the only trace we have of the discourse process. People make text out of common semantic resources so text provides us with essential data. The question is what evidence do these data provide of this discourse process of pragmatic meaning making and to answer that question we have to take into account the conditioning factors of context and pretext. So, to return to Alice once more, the question is not whether but how words can mean so many different things. Language is of its nature indeterminate, or it would not otherwise function pragmatically at all, so its meaning is always variable and always conditional. This, critically, is what critical discourse analysis actually reveals in spite of its assumption to the contrary. And this, I think, is the essential understanding we need to promote in language and literary education.

But now I must bring this talk to a close. My pretextual purpose, as will be obvious, has been to provoke you to think about the nature of text and how it gets interpreted. I must also admit to another pretext, which you will probably not be aware of: to encourage you to read a certain book, in which the points I have been making in this talk are dealt with more fully (Widdowson 2004).

A last point: everything I have been saying about text in general applies of course to the text I have been producing here today. As I have been producing it, you have been processing it-regulating your attention quite naturally according to your own pretexts, deriving no doubt different discourses from it. I do not know whether my intentions match up with your interpretations. Indeed, since all communication is partial and approximate, I shall never know how far what I have meant by my text corresponds with what it has meant to you.


Chomsky, Noam 2001, 9-11. New York: Seven Stories Press Fairclough, Norman 1992, Discourse and Social Change, Cambridge: Polity Press

Stubbs, Michael 1996, Text and Corpus Analysis, Oxford: Blackwell Sinclair, John 2004, Trust the Text: Language, Corpus and Discourse, edited by Ronald Carter, London/New York: Routledge van Dijk, Teun A. 2001, “Critical discourse analysis”, In The Handbook of Discourse Analysis, edited by Deborah Tannen, Deborah Schiffrin and Heidi Hamilton, Oxford: Blackwell, 352-371 Widdowson, Henry G. 2004, Text, Context, Pretext. Critical Issues in Discourse Analysis, Oxford: Blackwell

< Prev   CONTENTS   Source   Next >