Another important explanatory framework adopted in this book is the theory of social distinction drawn from the work of the French sociologist, Pierre Bourdieu. Bourdieu’s work, while admittedly drawn from his research on French society and relating largely to features of culture such as art, food and manners, nonetheless is relevant to language, as Bourdieu himself acknowledges. Bourdieu’s claim is a relatively simple one: features of culture are used to discriminate between groups in society, establishing a social hierarchy based on a series of social shibboleths.
The consequences of the establishment of such a hierarchy are both to allow members of groups to be readily identified and to impose the hierarchy itself. For example, if a taste for fine wine is supposed to be a token of high social status, then on seeing somebody pouring a drink from such a bottle of wine, other factors aside, one might assume they were of a certain social class.
Similarly, if one sees somebody drinking a pint of beer, and this is a marker of low social class, other factors aside, one may also infer their social class. However, if fine wine is priced so as to exclude the lower orders from purchasing it, the social hierarchy has nothing to do with taste as such. Rather, those tokens of taste are controlled in such a way as to impose the social structure that they are a token of. Transporting this argument to language is somewhat straightforward.
If there are forms of language which are identified with a refined form of speech, then those aware of the perception of this form of language, who are able to invest either the time or the money in order to acquire that ‘refined’ form of language, will be able to identify themselves with a particular group in society. Yet more perniciously, if that type of speech is already associated with a particular social class, then there is a zero cost for that social class in using that form of speech, while the speech associated with lower classes is devalued and the onus is placed on them to adapt the way that they speak.
In making that adaptation they are tacitly acknowledging the supposedly superior form of speech that they are shifting to when that shift takes place. To Bourdieu, in language this process leads to: opposition between popular outspokenness and the highly censored language of the bourgeois, between the expressionist pursuit of the picturesque or the rhetorical effect and the choice of restraint. 9 In seeking shibboleths of taste, groups distinguish themselves from one another in society in order to set boundaries which identify difference.
For Bourdieu this means that: Groups invest themselves totally, with everything that opposes them to other groups, in the common words which express their social identity, i. e. their difference. 10 In other words, the process of setting out the boundaries of linguistic differences for groups is no casual process. It is a process whereby the very identity of the groups concerned becomes intimately associated with their language use, through ‘the socially charged nature of legitimate language’.
11 Linked to a social hierarchy, the capacity is clearly generated to identify not merely the language of particular groups, but to identify the language of various groups with power as defining a discourse of legitimacy, a Swearing in English 10 discourse of power. This discourse of power then becomes the unmarked case—the linguistic norm, the supposedly neutral form of expression—with forms that do not follow it marked out as the marked, abnormal, negatively charged forms of language, or ‘the least classifying, least marked, most common, least distinctive, least distinguishing’12 forms of language.
This process of the discourse associated with one group becoming the dominant discourse of power leads to those not possessing that discourse being: at the mercy of the discourses that are presented to them… At best they are at the mercy of their own spokesmen, whose role is to provide them with the means of repossessing their own experience.
The essential indeterminacy of the relationship between experience and expression is compounded by the effect of legitimacy imposition and censorship exerted by the dominant use of language, tacitly recognized, even by the spokesmen of the dominated, as the legitimate mode of expression of political opinion. The dominant language discredits and destroys the spontaneous political discourse of the dominated. It leaves them only silence or a borrowed language.
13 In other words, those without access to this discourse of power are already marked as disadvantaged by their language use. This disadvantage is compounded by them having to use a discourse with which they do not readily identify when asserting themselves, as: Through the language… Bound up with a whole life-style, which foist themselves on anyone who seeks to participate in ‘political life’, a whole relation to the world is imposed. 14
At worst it may lead to the failure of the dominated groups to represent themselves, relying rather on members of the group possessing the dominant discourse consenting to represent them and provide leadership to them, as Bourdieu notes when he says that: It forces recourse to spokesmen, who are themselves condemned to use the dominant language…or at least a routine, routinizing language which…constitutes the only system of defence for those who can neither play the game nor ‘spoil’ it, a language which never engages with reality but churns out its canonical formulae.
15 Distinction simultaneously empowers further those already possessing power, while further dispossessing those who are already dispossessed. This book will argue that, when we look at modern English, we see distinction at work in the form of bad language. Broadly speaking, the discourse of power excludes bad language, the discourse of the disempowered includes it. Obviously, this statement is, however, something of an idealisation, as several factors may, for example, combine on any specific occasion to determine language usage.
Similarly, several factors together may establish a matrix of power, as opposed to single factors generating a polar distinction between the powerful and the disempowered. Indeed, in Part 2 of this book I will explore how demographic factors may combine in such a way. For the moment, I will maintain the broad assertion Bad language, bad manners 11 made above, adding the caveat that such a statement notes what is typical and is only generally applicable when we are considering one feature in isolation.
One final point I should make at this stage is that what I am discussing here is overt as opposed to covert prestige. 16 In this book I am mainly concerned with power related to overt prestige, though I accept without hesitation that in establishing an overtly prestigious form of language, a covertly prestigious form of language is entailed which may invert the matrix of power mentioned above. Research into overt and covert prestige is so well established that I feel the issue can be sidestepped in this book as there is a wealth of material that interested readers can pursue to explore this issue for themselves.
17 To recapitulate the earlier goals and claims of this book, it will be argued that the process of forming a class distinction around swear words was undertaken in the late seventeenth/early eighteenth century by an aspiring middle class who actively sought to distinguish themselves from the lower orders by a process of ‘purifying’ the speech of the middle class while prob-lematising the speech of the lower orders (see Chapters 4 and 6).
Further, it will be argued that the vehicle which brought about this process of distinction was a moral panic focused on bad language in the late seventeenth century, which empowered certain members of the middle classes to act simultaneously as moral entrepreneur and arbiter elegantium, dictating the linguistic manners of the general population. Finally, the book will argue that the processes of disempowerment which Bourdieu suggests are entailed by such a development are observable not merely in the seventeenth century but in the present day (see Chapters 5 and 7).
This brief overview of the book allows readers to see how the elements introduced in this chapter come together in order to provide a coherent account of bad language in English. The corpus is used principally to establish a series of observations related to distinctions in the use of swearing. The explanation of these distinctions is then sought through historical research, as well as the application of moral panic theory and Bourdieu’s theory of distinction to texts in the period 1690–1745. The process of disempowerment is then explored further in the context of debates about language in the media in the late twentieth century.
While corpus data will be instrumental in exploring the discourse of bad language in the seventeenth, eighteenth and twentieth centuries, it is the goal of this book to show that explanations for what we see in corpora often lie beyond the borders of the corpus itself—the observations we can draw from corpora, while verifiable, are not necessarily of any assistance in developing explanations, though they do frame what an acceptable explanation may look like, i. e. any explanation must match the observations drawn from the corpus.
But by marrying other methodologies with the corpus method, and drawing on appropriate theories, the corpus data itself can be illuminating in the search for a wider, comprehensive account of the features of language we approach the corpus to investigate. Corpus linguistics: the corpora used in this book In this section, I will discuss the majority of the corpora used in this book. Two minor corpora (a corpus of seventeenth-century news texts and a corpus of German radio propaganda broadcasts) will be discussed briefly when they are introduced.
One major Swearing in English 12 corpus, the Lancaster Corpus of Abuse,18 is not reviewed in this section, being reviewed instead in Chapter 2 as a prelude to an analysis of bad language in present-day English. The Mary Whitehouse corpus (MWC) The MWC includes the major writings of Mary Whitehouse in the period 1967–1977. This corpus covers three of her books, namely Cleaning-up TV, Who Does She Think She Is? and Whatever Happened to Sex? , amounting to 216,289 words in total.
19 These books, with their wide circulation, were the principal public output from the National Viewers’ and Listeners’ Association (the VALA—see Chapter 5 for details) in this period, and as such I take them to be a good focus for a study of how the VALA tried to excite a moral panic in the general population of Britain. The British National Corpus (BNC) The BNC is a 100,000,000-word corpus of present-day British English. The corpus is split into a 90,000,000-word balanced written corpus and a 10,000,000-word corpus of orthographically transcribed spoken language.
As I am using only the spoken data in this book I will limit my brief description of the BNC to its spoken section. The spoken BNC is composed of a series of spontaneous conversations recorded by members of the British public in the early 1990s. The corpus was designed to provide material from across the UK (the so-called demographically sampled subset of the corpus) and across a range of different activities (the so-called context governed subset). Demographic information about the speakers was encoded in the corpus.
This demographic data was then used to balance the spoken material with regard to a number of variables, notably, for this book, age, sex and social class. The result of this balancing is that, in the corpus, the amount of speech spoken by males and females is roughly even, as is the speech produced by different age groups and social classes. The Society for the Reformation of Manners corpus (SRMC) The SRMC was compiled by me specifically for this study. It contains four key texts from the Society for the Reformation of Manners (SRM) amounting to 120,709 words.
20 Two texts were selected as being those which achieved the widest circulation during the period of the reformation of manners movement and which were widely cited—Yates (1699) and Walker (1711)—while two further texts were included from the end of the period of the society’s activities, namely Anon. (1740) and Penn (1745). 21 The latter texts were included to permit an investigation of how, if at all, the discourse of the society shifted during its lifetime.
While ideally one would like to have gathered a much larger set of texts together, the longevity of the Yates and Walker texts, and their wide distribution during the lifetime of the societies, makes them in essence texts which are representative of the society and its aims. The later texts, as noted, represent some of the final texts of the society and are included solely to allow the possibility of a diachronic approach to the writings of the society. Bad language, bad manners 13 The Lampeter corpus The Lampeter corpus is a diachronic corpus of English, covering the period 1640–1740.
The corpus samples texts from a range of genres (economy, law, miscellaneous, politics, religion and science) over this period, taking samples at periods of roughly ten years. The corpus was constructed at the University of Chemnitz by a team led by Josef Schmied, and has been used in the diachronic study of variation in English. 22 For the purposes of this book, I will only use materials from the corpus covering the period 1690–1750, as it is in this period that I want to contrast the language of the SRM with what one might term English in general23 (i. e. all of the genres of the Lampeter at once) and specific genres and registers of English (texts covering only one domain of Lampeter). There are 544,894 words in the Lampeter corpus in the period 1690–1750. The Lancaster—Oslo—Bergen and Freiberg—Lancaster—Oslo— Bergen corpora (LOB and FLOB)
Both the LOB and FLOB corpora are related to an earlier corpus, the Brown University Standard Corpus of Present-day American English (i. e. the Brown corpus, see Kucera and Francis 1967). The corpus was compiled using 500 chunks of approximately 2,000 words of written texts.
These texts were sampled from 15 categories. All were produced in 1961. The components of the Brown corpus are given in Table 1. 1. LOB and FLOB follow the Brown model. The Lancaster—Oslo—Bergen corpus of British English (LOB) is a British match for the Brown corpus. 24 The corpus was created using exactly the same sampling frame, with the exception that LOB aims to represent written British English used in 1961. The Freiberg-LOB corpus of British English (i. e. FLOB) represents written