The study of political discourse, like that of other areas of discourse analysis, covers a broad range of subject matter, and draws on a wide range of analytic methods. Perhaps more than with other areas of discourse, however, one needs at the outset to consider the reflexive and potentially ambiguous nature of the term political discourse. The term is suggestive of at least two possibilities: first, a discourse which is itself political; and second, an analysis of political discourse as simply an example discourse type, without explicit reference to political content or political context. But things may be even more confusing. Given that on some definitions almost all discourse may be considered political (Shapiro 1981), then all analyses of discourse are potentially political, and, therefore, on one level, all discourse analysis is political discourse.
This potentially confusing situation arises, in the main, from definitions of the political in terms of general issues such as power, conflict, control, or domination (see Fairclough 1992a, 1995; Giddens 1991; Bourdieu 1991; van Dijk 1993; Chilton and Schaffer 1997), since any of these concepts may be employed in almost any form of discourse. Recently, for example, in a study of a psychotherapeutic training institution, Diamond (1995) refers to her study of the discourse of staff meetings as “political,” simply because issues of power and control are being worked out. They are being worked out at different levels, however: at interpersonal, personal, institutional, and educational levels for example, and in different strategic ways (Chilton 1997). By treating all discourse as political, in its most general sense, we may be in danger of significantly overgeneralizing the concept of political discourse.
Perhaps we might avoid these difficulties if we simply delimited our subject matter as being concerned with formal/informal political contexts and political actors (Graber 1981); with, that is, inter alia, politicians, political institutions, governments, political media, and political supporters operating in political environments to achieve political goals. This first approximation makes clearer the kinds of limits we might place on thinking about political discourse, but it may also allow for development. For example, analysts who themselves wish to present a political case become, in one sense, political actors, and their own discourse becomes, therefore, political. In this sense much of what is referred to as critical linguistics (Fairclough 1992b) or critical discourse analysis (van Dijk 1993; Wodak 1995) relates directly to work on political discourse, not only because the material for analysis is often formally political but also, perhaps, because the analysts have explicitly made themselves political actors (see van Dijk, this volume).
But such a delimitation, like all delimitations, is not without its problems. For example, how do we deal with the work of Liebes and Ribak (1991) on family discussions of political events? Is this political discourse, or family discourse of the political? In one sense it is both – but the issue of which may simply be a matter of emphasis (see, for example, Ochs and Taylor 1992). While delimitations of the political are difficult to maintain in exact terms, they are nevertheless useful starting points. Equally, while one can accept that it is difficult to imagine a fully objective and nonpolitical account of political discourse, analysts can, at best, and indeed should, make clear their own motivations and perspectives. This may range from setting some form of “democratic” ideal for discourse against which other forms of political discourse are then assessed (Gastil 1993) to explicitly stating one's political goals in targeting political discourse for analysis (as in the case of a number of critical linguists: Fairclough 1995; Wodak 1995; van Dijk 1993). It also allows for more descriptive perspectives (Wilson 1990, 1996; Geis 1987), where the main goal is to consider political language first as discourse, and only secondly as politics. The general approach advocated above would respond to the criticism of Geis (1987), who argues that many studies of political language reveal their own political bias. Most of us who write about political discourse may do this at some level, but as long as this is either made clear, or explicitly accepted as a possibility, then this seems acceptable.
The study of political discourse has been around for as long as politics itself. The emphasis the Greeks placed on rhetoric is a case in point. From Cicero (1971) to Aristotle (1991) the concern was basically with particular methods of social and political competence in achieving specific objectives. While Aristotle gave a more formal twist to these overall aims, the general principle of articulating information on policies and actions for the public good remained constant. This general approach is continued today.
Modern rhetorical studies are more self-conscious, however, and interface with aspects of communication science, historical construction, social theory, and political science (for an overview see Gill and Whedbee 1997). While there has been a long tradition of interest in political discourse, if one strictly defines political discourse analysis in broadly linguistic terms (as perhaps all forms of discourse analysis should be defined: see Fairclough and Wodak 1997), it is only since the early 1980s or 1990s that work in this area has come to the fore. Indeed, Geis (1987) argues that his is the first text with a truly linguistic focus on political language/discourse. There is some merit in this argument, but without opening up issues about what is and what is not linguistics, many of the earlier studies in social semiotics and critical linguistics should also be included in a general linguistic view of political discourse (Fowler et al. 1979; Chilton 1990, 1985; Steiner 1985). While language is always clearly central to political discourse, what shifts is the balance between linguistic analysis and political comment. Distinguishing the direction of this balance, however, is not always straightforward.
In more modern times it was perhaps Orwell who first drew our attention to the political potential of language. This is seen in his classic article “Politics and the English Language,” where he considers the way in which language may be used to manipulate thought and suggests, for example, that “political speech and writing are largely the defence of the indefensible” (1969: 225). His examples are types of inverted logic (reflected in literary detail in his book Nineteen Eighty Four) and they echo through much of the present work on political discourse. Instances include the use of “pacification” to refer to the bombing of defenseless villages, or the use of “rectification of frontiers” to refer to the relocation or simply removal of thousands of peasants from their homes.
Orwell was concerned with a general decline in the use of English, and politicians had a central responsibility for this decline. They have a general reputation for the construction of what Americans call “fog” or the British political gobbledygook (see Neaman and Silver 1990: 320). For example, the American navy have described high waves as “climatic disturbances at the air-sea interface,” while in the 1970s, President Nixon's press secretary coined the phrase “biosphere overload” for overpopulation (also called “demographic strain” by some government officials) (see Neaman and Silver 1990: 317–21). The British are not exempt from such excesses of lexical production, however; an antivandalism committee of the Wolverhampton District Council was given the title, “The Urban Conservation and Environmental Awareness Work Party” (Neaman and Silver 1990: 321).
However, it is not simply manipulation that is at issue in the case of political language; it is the goal of such manipulation which is seen as problematic. Politicians seem to want to hide the negative within particular formulations such that the population may not see the truth or the horror before them. This is the general thrust of Orwell's comments, and it emerges again and again throughout work on political discourse, but with perhaps different levels of emphasis and analysis. The influential work of the political scientist Murray Edleman (1971, 1977, 1988) mirrors Orwell's concerns and looks at the symbolic manipulation of reality for the achievement of political goals. In a more directed political sense Pěcheux (1982, 1978), following Althusser's claim that ideology is not just an abstract system of thought but becomes actualized in a variety of material forms, set about studying discourse as one type of material form. Pěcheux argued that the meanings of words became transformed in terms of who used them, or, in Foucault's (1972) terms, in relation to particular “discourse formations.” Here words (and their interaction) in one formation were differently interpreted within another. Conservative or right-wing views of terms like “social benefit” and “defense spending” may differ radically from interpretations available within a socialist or left-wing discourse (see below).
The general principle here is one of transformation. Similar words and phrases may come to be reinterpreted within different ideological frameworks. Linked directly to this process is the concept of “representation.” Representation refers to the issue of how language is employed in different ways to represent what we can know, believe, and perhaps think. There are basically two views of representation: the universalist and the relativist (Montgomery 1992). The universalist view assumes that we understand our world in relation to a set of universal conceptual primes. Language, in this view, simply reflects these universal possibilities. Language is the vehicle for expressing our system of thought, with this system being independent of the language itself. The relativist position sees language and thought as inextricably intertwined. Our understanding of the world within a relativist perspective is affected by available linguistic resources. The consequences here, within a political context, seem obvious enough. To have others believe you, do what you want them to do, and generally view the world in the way most favorable for your goals, you need to manipulate, or, at the very least, pay attention to the linguistic limits of forms of representation.
While many analysts accept the relativist nature of representation in language, i.e. that experience of the world is not given to us directly but mediated by language, there is a tendency to assume that politically driven presentation is in general negative. In Fairclough's (1989) view of critical linguistics/discourse, for example, political discourse is criticized as a “form of social practice with a malign social purpose” (Torode 1991: 122). The alternative goal is “a discourse which has no underlying instrumental goals for any participant, but is genuinely undertaken in a co-operative spirit in order to arrive at understanding and common ground.”
Examples of this malign social purpose are highlighted in work on the political discourse of what has been referred to as “nukespeak.” As is clear, the very title “nukespeak” is formed on analogy with Orwell's famous “newspeak,” where the assumption was that if one could manipulate or limit what was possible in language then one could manipulate or limit what was possible in thought. Chilton (1985) and others argue, using a range of analytic techniques, that in the political discourse of nuclear weapons efforts are made to linguistically subvert negative associations. An example from Montgomery (1992: 179) highlights this general issue (see also Moss 1985):
Strategic nuclear weapon – large nuclear bomb of immense destructive power
Tactical nuclear weapon – small nuclear weapon of immense destructive power
Enhanced radiation weapon – neutron bomb (destroys people not property)
Demographic targeting – killing the civilian population
In this example Montgomery is performing a type of translation in which he explicitly attempts to show how the language on the left of the dash is manipulating reality as represented by the translation on the right. For Montgomery, the language of nuclear weapons is clearly “obscurantist and euphemistic.”
A similar and related point to that noted in Montogmery's work has been made specifically in the case of syntax (Montgomery 1992; Simpson 1988, 1993; Chilton 1997). The system of “transitivity,” for example (Halliday 1985), provides a set of choices for describing “what is going on in the world.” One such choice is referred to as a “material process,” where what is going on may be described as an action, transaction, or event.
An example from Goodman (1996: 56) clearly illustrates these options:
|a.||The solider (Actor)||fired (material process: action)|
|b.||The soldier (Actor)||killed (material process: transaction)||innocent villagers (goal)|
|c.||Innocent villagers (goal: material process)||died (material process: event)|
Goodman (1996:57) comments on the possible reasons behind such selections, suggesting:
Writers with a technical interest in weaponry (in a specialist magazine) might have an interest in obscuring the pain and destruction that weapons cause. Writers who are on the same side as the soldiers might also have an interest in obscuring their army's responsibility for the death of innocent civilians.
Although Goodman is writing in 1996, we can note the similarity with Orwell's comments some 50 years previously (see also Chilton 1997; Stubbs 1996). While many of Goodman's claims may be true, Fairclough (1995) notes that such claims are often built around single, isolated utterances, taking no account of the textual or historical context of production. One might, for example, decide to present the sentences highlighted by Goodman by sequencing the events for the listener in very specific ways:
Innocent villagers died last night. It was the soldiers who fired on them. It was the soldiers who killed them!
In the first sentence here it is the villagers who are highlighted, not the soldiers. One might argue, as does Goodman, that such a form obscures those responsible. However, not only are those responsible highlighted in the next two sentences, but the very contrast that is indicated by their exclusion from the first and not the following sentences might lead readers back to the first sentence to confirm their originally hidden responsibility. By inviting readers/listeners to revisit the first sentence, this small text may emphasize not only the responsibility of the soldiers, but that they have tried to avoid that responsibility.
Issues of representation, however, need not only revolve around specific syntactic transformations: without any seemingly manipulative intent one can achieve personal and political goals by relatively uncontroversial structural selections. Consider the general area of evidentiality. Evidentiality refers to the way in which forms of evidence become grammaticalized in different languages and to the attitude one takes or adopts toward this evidence (see papers in Chafe and Nichols 1986), since not all evidence is of a similar type. There is a complex interaction here between such things as beliefs, assumptions, inferences, and physical experiences (sight, hearing, smell, touch, etc.): I saw John yesterday; I believe I saw John yesterday; I was told John was seen yesterday; it is possible that John was seen yesterday.
In a study of political discourse just prior to American entry into the 1990 Gulf War, Dunmire (1995) argues that newspaper articles in both the New York Times and the Washington Post, and statements made by representatives of the American government, actively assisted the USA in positioning itself for intervention. They did this by shifting their concerns from Iraq's invasion of Kuwait to a series of claims regarding a potential attack on Saudi Arabia. Dunmire argues that, through an increased use of nominal clauses to represent the threat of Iraq's attack on Saudi Arabia, what was speculation came to be accepted as fact.
Equally, it may be that in some cases it is not simply the syntactic form which is chosen, but rather the relative distribution of particular syntactic selections which carries the political implications. Work by Stubbs (1996) on the distribution of ergative forms within two school geography textbooks may be used to illustrate this point. As Stubbs (1996: 133) explains, ergatives are verbs which: can be transitive or intransitive, and which allow the same nominal group and the same object group in transitive clauses and as subject in intransitive clauses: several firms have closed their factories factories have been closed factories have closed
The important point is that ergatives have agentive and nonagentive uses. This allows ergatives, like transitivity in active and passive sentences, to be used differentially depending on the ideological goals of the text.
Using a computer analysis of two different types of school text, one which looked at human geography from a fact-based perspective (text G), and one which adopted an environmentalist position (text E), Stubbs discovered significant distributional differences between the two:
Relative to text length texts G and E have almost the same number of ergative verbs: slightly fewer than one per 100 words of running text. However, the distribution of transitive, passive, and intransitive choices is significantly different (p < 0.001). Text E has many more transitive forms with correspondingly fewer passives and intransitives. Consistent with explicit orientation to the responsibility for environmental damage, Text E expresses causation and agency more frequently. (Stubbs 1996: 137)
Clearly, text E's author has adopted an explicit political role within the text and this is revealed through both a grammatical and a distributional analysis of specific verb forms.
The idea that similar grammatical categories may be operationalized in different ways is taken up by Kress and Hodge (1979), who have argued that several different types of strategy might be subsumed under a general heading of negation. They explore the use of a variety of options available to politicians which allow them to articulate some contrastive alternatives to what they are saying: I agree with you but…; that is a fair point, nevertheless…; I see your point yet…. However, such stylistic assumptions seem to overlap with other levels of structure such as discourse, for example, and indeed forms such as but, nevertheless, well, etc. are now normally referred to as discourse markers (Schiffrin 1987; Gastil 1993; see Schiffrin, this volume). Wilson (1993) explicitly treats such forms as discourse markers and suggests that they may function differentially in the marking of ideological contrasts. In an analysis of students' debates on specific political subjects, it is noted that “and” may be used for either planned coordination (as in X, Y, and Z) or unplanned coordination (as in X and Y and Z). The choice one adopts relates to the way one wishes to present the elements coordinated by “and.” In political terms, unplanned coordination is used where one wishes the elements to be treated independently (Scotland and England and Wales and Northern Ireland), whereas planned coordination treats the elements as naturally linked (Scotland, England, Wales, and Northern Ireland).
Linguistic options for representing the world are clearly, then, central issues in political discourse, but so are issues of action and textual production. Utterances within the context of political output are rarely isolated grammatical cases; they operate within historical frameworks and are frequently associated with other related utterances or texts (Bakhtin 1981). In 1993, for example, the prime minister of Britain responded to a question in the House of Commons in the following way:
PM John Major: “If the implication of his remarks is that we should sit down and talk with Mr. Adams and the Provisional IRA, I can say only that would turn my stomach and those of most hon. Members; we will not do it. If and when there is a total ending of violence, and if and when that ending of violence is established for a significant time, we shall talk to all the constitutional parties that have people elected in their names. I will not talk to people who murder indiscriminately”. (Hansard Official Report, November 1, 1993: 35)
Despite this statement, however, on November 15, 1993, Gerry Adams, the leader of Sinn Fein, claimed that the British government was, in fact, involved in protracted dialogue with Sinn Fein. The claim was rejected by the British government, but Adams went on to claim that Major had broken off the contact “at the behest of his Unionist allies” (Belfast Telegraph, December 15, 1993). The next day Sir Patrick Mayhew, the secretary of state for Northern Ireland, when asked on BBC Television if there had been contact with Sinn Fein or the IRA by people who could be regarded as emissaries or representatives of the British government, said “No there hasn't.” The controversy over government contacts with the IRA resurfaced when, on November 22, Mayhew announced that “Nobody has been authorised to talk to or to negotiate on behalf of the British Government with Sinn Fein or any other terrorist organisation” (Belfast Telegraph, December 15, 1993). However, reports in the Observer newspaper later that week forced the government to admit having been in contact with the IRA in response to an IRA peace overture in February of that year.
Both journalists and Unionist politicians were by now beginning to argue that at best the government had misled them, and at worst lied to them (see Ian Paisley's comments in Hansard, November 29, 1993: 786). The government insisted that any contact had been at arm's length. On November 28 Sir Patrick admitted that the meetings had been going on for three years. The following day in the Commons he was forced to account for the seeming discrepancy between government statements and government actions.
The general claims made by Mayhew in the House of Commons were summarized and paraphrased in Wilson (1993: 470) as follows:
· We did not (1a) talk to the IRA, we had channels of communication/contacts.
· We did not (1b) authorise anyone to talk with the IRA.
In the first case a semantic contrast between talk and communication is presented, the claim seemingly being that the British government did not have articulate verbal contact, but did communicate with the IRA using selected channels of communication. In (1b) negation is employed in the context of a particular type of presuppositional verb (authorize) which creates two possible interpretations, both of which are equally acceptable:
We did not authorize anyone to talk to the IRA, so no one did.
We did not authorize anyone to talk to the IRA, although someone did (unauthorized).
Which statement was intended was never made clear in the debates that took place. However, as a number of politicians indicated at the time, the issue was not whether the government had communication channels with the IRA, but that John Major (and the secretary of state in other statements) implied by their comments (“[to] talk with Mr. Adams and the Provisional IRA… would turn my stomach”) that the British government would not have any contact with the IRA until they gave up violence. For some of the politicians who listened to John Major's original claims, any contact at whatever level, authorized or unauthorized, was in breach of such claims.
This particular incident involves a complex of textual and historical issues as well as examples of particular forms of representation. It illustrates the need for arguments about political manipulation to draw on larger-scale linguistic structures, as well as general grammar and single words or phrases.
This is not to deny the significance of single words or phrases in the discussion of political discourse; the aim is merely to highlight other relevant aspects in delimiting political discourse. But even at the level of words and phrases themselves, as Stubbs has shown, it may not merely be the single occurrence of a term that is important but sets of collocational relationships, which in their turn produce and draw upon ideo logical schemas in confirming or reconfirming particular views of the world. For example, Stubbs (1990: cited in Stubbs 1996: 95) analyzed a newspaper text of riots in South Africa and showed how blacks and whites were frequently described by differ ent sets of words (see Wodak and Reisigl, this volume):
Blacks act in mobs, crowds, factions, groups. They constitute millions, who live in townships and tribal homelands. They mass in thousands and are followers of nationalist leaders. But Whites (who are also reported as committing violence) are individuals or extremists. By implication different from other (normal) Whites.
On a related level there is a further potential problem with some of the examples of political representation noted above, and this is that relativism affects everyone, including the analyst. The descriptive and, indeed, manipulative element in ana lyses concerned with the way in which representation may become systemically struc tured for specific effect is not in doubt. The derived implications may sometimes, however, be more political than analytical. At one level there is a suggestion that heroic terms for weapons, such as tomahawk, peacekeeper, Hawkeye, etc. (Moss 1985: 56), or the reordering of events (active vs. passive), reconstitute the world for hearers such that the truth or reality of an event is subverted. I have no doubt of the gen eral truth in this, but along with Horkheimer (1972) and Garfinkel (1967), I do not view participants to communication as potential “interactional” dopes but rather, as Giddens (1991) suggests, social actors capable of making choices, no matter how constrained the conditions. As Giddens notes, an agent who has no choice is no longer an agent.
Equally, since the transitive system of English syntax is available to all Eng lish speakers, alternative ways of representing the world may not be interpreted by hearers in exactly the ways that producers intend. As suggested above, the transforma tion of a passive sentence in production into an active sentence in interpretation is perfectly feasible. Indeed, research into political information processing clearly indic ates that interpretation in affected by cognitive bias (St Evans 1989). Once information is encoded into memory in terms of one set of concepts, it is unlikely to be retrieved and interpreted in terms of other, alternative sets presented at a particular point in time. For example, people who have conceptualized their view of blacks in a particular negative way are unlikely to adjust that view on reading or hearing a text which has manipulated any presentation of this group in a more positive manner. This does not suggest there are no possibilities for change, however. Views can be reformulated given forms of counterevidence presented over time and brought forward in parti cular ways, and part of this reformulation will, of course, be through different lin guistic presentations. The fact is, however, that specific biases may override structural presentation.
This may be seen clearly in attempts to model ideological reasoning in a computa tional form. One of the best known systems is POLITICS (see Carbonell 1978; see also Hart 1985), which is a program designed to interpret political events in relation to differing ideological frames. For example, if the input is (2), then the output for a conservative interpretation of the event would be (3) and that for a liberal interpreta tion of the same event would be (4):
· The United States Congress voted to fund the Trident Submarine project. (2)
|(3)||a.||The United States Congress wants the United States armed forces to be stronger,|
|b.||The United States Congress should be strong to stop communist expansion.|
|(4)||a.||The United States Congress fears falling behind in the arms race.|
|b.||The United States should negotiate to stop the arms race, (adapted from Carbonell 1978:30)|
The reference to an arms race or communist threat dates the POLITICS system. The important point nevertheless is that such systems generally work on the basis of key propositions within the input. These are then linked to particular scripts or frames (Schank and Ableson 1977); for instance, what the USA should do in the case of nuclear threat. These scripts provide a mechanism for grouping inferences and de fining the context in which interpretation takes place. Such contexts are modified relative to certain ideological formations (conservative or liberal). While it would be possible to build in specific parsing constraints which may be sensitive to structural dimensions of syntax, the important features for the system are elements such as “Congress” and “fund,” not necessarily their syntactic embedding.
As suggested above, syntactic selection undoubtedly affects interpretation, but this must be seen in relation to other contextual factors, and indeed in relation to the impact of lexical choices themselves. Wilson and Rose (1997) argue, for example, that the problems of interpretation which accompanied one piece of controversial legisla tion, the 1985 Anglo-Irish Agreement, seemed to revolve around single lexical items. Making use of Sperber and Wilson's (1996) theory of relevance, Wilson and Rose describe how a single lexical item, in this case consultation, drives differing interpreta tions of the agreement. This controversial legislation brought together the Irish and British governments in an intergovernmental forum. The British government described the relationship as one of “consultation,” and modified this as “merely consultation,” revealing their view that they were only talking to the Irish government as opposed to being influenced by them. The Irish government, in contrast, viewed “consulta tion” as a process of influence. One does not normally consult someone unless one is willing to take the person's advice. In this case, consultation meant more than discussion; it was discussion plus impact. This was also the interpretation given by the Unionist parties within Northern Ireland, who were vehemently opposed to the agree ment. On the other hand Sinn Fein accepted the British interpretation, and for this very opposite reason (i.e. the British would do nothing more than talk to the Irish government) they also opposed the agreement. The point is, however, that in the myriad debates which took place at the time, the syntax of presentation seemed to have little impact on ideologically contrived lexical interpretations.
Such conflicts over lexical interpretation are not new, of course. Everyday words, organized and structured in particular ways, may become politically implicated in directing thinking about particular issues, and with real and devastating effects. Even the process of uttering someone's name may become a political act, as it did in the infamous McCarthy trials of the 1950s (see also Wilson 1990: ch. 3).
McCarthy's witch-hunt for communists created a context where “naming names” became a central issue (see Navasky 1982). The McCarthy trials raised questions about the very act of naming and what it means to name someone in certain kinds of social context. If one agreed to name names, was one an informer” or an inform ant,” for example? Ultimately, this depended on which side of the semantic fence you stood on. J. Edgar Hoover, the head of the Federal Bureau of Investigation, was quite clear on his position:
They stigmatize patriotic Americans with the obnoxious term “informer,” when such citizens fulfil their obligations of citizenship by reporting known facts of trie evil conspiracy to properly constituted authorities. It would require very little time for these critics to pick up a dictionary; Webster's unabridged volume specifically states that an “informant” is one who gives information of whatever sort; an in former is one who informs against another by way of accusation or complaint. Informer is often, informant almost never, a term of opprobrium, (cited in Navasky 1982: xviii)
Whatever one's reasons for providing names to McCarthy's committee – and Navasky notes that justification ranged from the protection of the country (where one Manning Johnson admitted he would lie in a court of law in the course of protecting his country) to liberal outrage (James Wechsler argued that only by cooperating with the committee could he gain access to a transcript of the trials, which he could then use to attack the committee itself) – in those cases where names were provided a number of analysts took a simpler and alternative view to Hoover's: Navaksy (1982) states quite straightforwardly that anyone who gave names “was an informer.”
The interesting issue in all this is in relation to what one believes a word means, and what effect, beyond a word's core or semantic meaning, the use of the word has. Hoover objected to the use of the word “informer” not because it cannot be, in one sense, correctly applied to anyone who gives names, but because it carries negative connotations, and he believed that the actions of naming within the context of the search for communists and communist sympathizers ought to be seen as positive. Navasky takes an opposing view; despite Hoover's suggested semantic arguments, he points out that most of those who gave evidence thought of themselves as infor mers, and, says Navasky, “that's what I will call them” (1982: xviii).
Or consider another context where ordinary, everyday words are organized differ ently within the discourse of speechmaking. The following extracts are taken from a speech given by Neil Kinnock, at the time the Labour Opposition leader in Britain, on Tuesday June 2, 1987, at a Labour Party rally in Darlington, England: Uunemployment is a contagious disease … it infects the whole of the economic body …If limbs are severely damaged the whole body is disabled. If the regions are left to rot, the whole country is weakened …… just as the spread of unemployment, closure, redundancy, rundown … affects the economic life in that region so the same ailments in a country gradually stain the whole country.… if the battered parts and people of Britain don't get noisy they will just get neglected. Silent pain evokes no response.
What is clear from these extracts, and many others within the same speech, is that the semantic fields of illness and health are being evoked in an attempt to produce rel evant political images. Some of the vocabulary employed in this effort is highlighted below:
Fracture, illness, decay, deprivation, contagious, (contagious) disease, body, strength, (shrivel), cuts, limbs, damage (severe), disabled (body), weakened, spread (disease), rundown, ailments, battered (parts), pain, dose (decline), deaden, waste, accident, healing, caring, disabled, short-sighted, welfare, chronically ill, affliction, handicapped, medicine, infects
It is also clear that many of these terms are negatively marked. Examples are weak as opposed to strong; dead as opposed to alive; decline as opposed to revival; and ill as opposed to well. It would, of course, be possible for Kinnock to use these terms to actually refer to the health issues of real groups of people, and within the speech the use of handicapped would fall into this category. Nevertheless, the majority of words taken from the area of health (see below) are employed out of context, that is, in this case, metaphorically.
This is a further reflection of Fairclough's (1995) general point about not looking at isolated sentences, or in this case isolated words. While much has been made of single words in political discourse (Wodak 1989; Hodge and Fowler 1979; Geis 1987; Bolinger 1982), the reality is that in most cases it is the context, or reflected form (Leech 1995), of the words which carries the political message. This is particularly true of the kinds of metaphorical uses made by Kinnock. As Lakoff and Johnson (1980) have shown, metaphorical uses may describe the world for us in particular ways such that we come to understand the world in that way (representation again: see Chilton and Ilyin 1993). And this is what Kinnock is trying to do. What he wants is for us to understand the world in such a way that all aspects of Conservative government control lead to disease and decay.
The issue here, as with both the POLITIC system interpretation and the human interpretation of the Anglo-Irish Agreement, is that some humans, like some systems, may be biased in their mode of interpretation from the start. For such individuals, manipulations of transitivity, or other aspects of structure, may have little effect on interpretation, which is not to say that such structural forms may not have an impact elsewhere. The point is that there are many dimensions of language involved in political output, and all of these have the potential in their own way for political impact. Even individual sounds may become political, and a much-neglected area of political language is what we might call “political phonology.”
It may be initially difficult to grasp how specific sounds come to be interpreted as political, although where one sees politics as tied directly to forms of ideology, the issue becomes a central plank of variationist sociolinguistics, and beyond (see Cameron 1995; Lippi-Green 1997). Research on accent clearly indicates that selected phonolo gical variables can carry political loading. By their very nature, phonological vari ables have been tied to issues such as class, gender, and ethnicity, and, in turn, to the social and political implications of the use of such variables (at both macro- and microlevels; Wilson and O Brian 1998).
Despite this natural link between phonological work in variationist sociolinguistics and political and social facts, there have been few studies of the potential of phono logy in the direct construction of political discourse. There is no reason to presuppose, however, that this level of linguistic structure may not also be available for political orientation. There is general evidence, for example, that Margaret Thatcher modified her speech in very particular ways in order to make herself more attractive to voters. And in the work of Gunn (1989; Wilson and Gunn 1983) it is claimed that leading politicians and political supporters may make adjustments within their phonological systems for political effect. For example, Gerry Adams is said to have adopted phono logical forms as representative of southern Irish dialect alternatives, and placed these within his own Belfast phonological system. Similarly, selected members of the Democratic Unionist Party, at the opposite end of the political spectrum from Adams's Sinn Fein, were shown to modify some of their phonology in the direction of a perceived and geographically (North Antrim) located Ulster Scots dialect. What this means is that politicians can choose to sound ideological/political, and indeed that such modifications are perceptually salient to the public. Matched guise studies (see Lambert et al. 1960), manipulating the kinds of phonological variables noted by Gunn (Wilson and Gunn 1983), revealed that certain variables were associated with political factors such as Unionism and Republicanism and general social factors such as Protest antism, Catholicism, Britishness, and Irishness. By adopting particular alternative phonological forms, one could be perceived as either more Catholic/Irish/Republican or more Protestant/British/Unionist.
One of the core goals of political discourse analysis is to seek out the ways in which language choice is manipulated for specific political effect. In our discussions we have clearly seen that almost all levels of linguistics are involved; i.e. most samples of political discourse may be mapped onto the various levels of linguistics from lexis to pragmatics. At the level of lexical choice there are studies of such things as loaded words, technical words, and euphemisms (Graber 1981; Geis 1987; Bolinger 1982). In grammar, as we have seen, there are studies of selected functional systems and their organization within different ideological frames (Fowler and Marshall 1985). There are also studies of pronouns and their distribution relative to political and other forms of responsibility (Maitland and Wilson 1987; Wilson 1990; Pateman 1981; Lwaitama 1988) and studies of more pragmatically oriented objects such as implic- atures, metaphors, and speech acts (van Dijk 1989; Wilson 1990; Holly 1989; Chilton and Ilyin 1993).
As we have discussed above, defining political discourse is not a straightforward matter. Some analysts define the political so broadly that almost any discourse may be considered political. At the same time, a formal constraint on any definition such that we only deal with politicians and core political events excludes the everyday discourse of politics which is part of people's lives. The balance is a difficult one, and perhaps all we can expect from analysts is that they make clear in which way they are viewing political discourse, because they too, like politicians, are limited and mani pulated in and by their own discourse. As we have seen, in a number of cases (Stubbs and van Dijk, for example) the text which is being analyzed has already been delimited as a specific political type. Stubbs refers to his chosen text as an “environmentalist one,” and van Dijk refers to specific speeches as “racist.” In both cases, social and political judgments have been made before analysis commences. In other studies (Gunn and Wilson, for example) the data generate their own stories, and the initial constraint is usually only linguistic, the political being drafted in later to explain why patterns may have emerged as they have. I am not suggesting that these are mutually exclusive alternatives, or that one or the other has any specific problems. The point is made to illustrate the way in which some analyses may become as much political as linguistic; and I think political discourse is made up of, and must allow for, both.
Since the early 1980s, there has been a growing interest in the area of political discourse (with studies emerging from across the globe: see Chilton 1997). While many studies have adopted (explicitly or implicitly) a critical perspective (see van Dijk, this volume), there has also been a variety of other approaches available, rang ing from the descriptive to the psychological. The essential issue in political discourse is, as we have noted, the balance between linguistic analysis and political analysis, and we have perhaps emphasized the former in this chapter as opposed to the latter, since, in general, this is what distinguishes political discourse analysis from political research as found, say, in political science.
It is also now a growing trend in political discourse to combine social theory with linguistic theory (see Fairclough 1992a; Wodak 1995). The trick, however, is not to lose linguistic rigor for the sake of sociopolitical claims, but equally not to simply continue producing language-based analyses which do not fully consider why, in social and political terms, specific linguistic choices have been made. There is also an emerging argument for a more integrated semiotic view of public and political com munications which combines analyses of a range of sign-based systems (Kress and van Leeuwen 1990, 1996). But certain core features will, and must, remain constant in the field of political discourse, and central to this is the role of language and lan guage structure, and its manipulation for political message construction and political effect.
1. Aristotle. (1991) On Rhetoric: A Theory of Civil Discourse, trans G. Kennedy . Oxford: Oxford University Press.
2. Bakhtin, M. (1981) The Dialogical Imagination, ed. M. Holquist, trans. C. Emerson and M. Holquist. Austin: University of Texas Press.
3. Carbonell, J. G. (1978) POLITICS: automated ideological reasoning. Cognitive Science (2) , 27–51.
4. Chilton, P. (1985) Words, discourse and metaphors: the meanings of deter, deterrent and deterrence. In P.
5. Chilton, P. (1987) Discourse and politics. In T. van Dijk (ed.), Discourse as Social Interaction. London: Sage, 206–30.
6. Chilton, P. (1990) Politeness and politics. Discourse and Society (1) (2), 201 24.
7. Chilton, P. and Ilyin, M. (1993) Metaphor in political discourse. The case of the common European house. Discourse and Society (4) (7), 7–31.
8. Diamond, P. (1995) Status and Power in Verbal Interaction: A Study of Discourse in a Close-knit Social Network Amsterdam: Benjamins.
9. Edleman, M. (1977) Political Language New York: Academic Press.
10. Edwards, D. and Potter, J. (1992a) Discursive Psychology London: Sage.
11. Fairclough, N. (1989) Language and Power London: Longman.
12. Garfinkel, H. (1967) Studies in Ethnomethodology Englewood Cliffs, NJ: Prentice-Hall.
13. Gastil, J. (1993) Undemocratic discourse: a review of theory and research on political discourse. Discourse and Society (3) (4), 469–500.
14. Geis, M. (1987) The Language of Politics New York: Springer Verlag.
15. Giddens, A. (1991) Modernity and Self Identity Cambridge: Polity.
16. Gill, A. M. and Whedbee, K. (1997) Rhetoric. In T. van Dijk (ed.), Discourse as Structure and Social Process. London: Sage, 157–85.