Sociolinguistics PDF

Chapter 9 Sociolinguistics Carmen Llamas1 and Peter Stockwell2 What is sociolinguistics? The most obvious definition of ‘sociolinguistics’ is that it is the study of language in society. However, if it was as easy as that, then almost every language event would form part of the field of sociolinguis...

Chapter 9 Sociolinguistics Carmen Llamas1 and Peter Stockwell2 What is sociolinguistics? The most obvious definition of ‘sociolinguistics’ is that it is the study of language in society. However, if it was as easy as that, then almost every language event would form part of the field of sociolinguistics. After all, there is a social and contextual dimension to every naturally occurring use of language, and it is always these social factors that determine the choice and form of what is written or said or understood. If sociolinguistics is not to encompass all linguistics, psychology and social theory, then we need a more precise and complex definition. So, sociolinguistics is the study of the linguistic indicators of culture and power. This is much more specific. This allows us to focus on language but also allows us to emphasize the social force of language events in the world. It allows us to use the tools of linguistics as outlined in the first part of this book (grammar, vocabulary, corpus linguistics, discourse analysis and pragmatics), as well as phonology, but it also encourages us to see the influences of ethnicity, gender, ideology and social rank on language events. Above all, this definition allows sociolinguists to be descriptive of pieces of language in the world, whilst encouraging us to recognize that we are all included in that world too. It could even be argued that sociolinguists have a special responsibility to use their privileged knowledge to influence the direction of, for example, government language policies, educational practices, media repre- sentations and so on. Many sociolinguists have argued strongly for this ethically-involved position. How- ever, we must recognize that the majority of sociolinguistic studies are primarily descriptive and aim towards a scientific objectivity, even when dealing with very com- plex social influences on language. That is, most studies focus on giving an account of social aspects of language in the real world that is as precise and systematic an account as possible within the current state of knowledge. Sociolinguistics is thus progressive as a discipline in the sense that new studies and new thinking are continually testing and developing our understanding of the way language and society work in relation to each other. This means we need a definition of sociolinguistics that covers the central con- cerns of the majority of the discipline. So, finally and centrally, sociolinguistics is the study of language variation and lan- guage change. This definition foregrounds the essential features of language: societies differ from each other and change over time, and language is bound up with these processes. The two dimensions can be seen as complementary axes: an historical or ‘diachronic’ axis which is concerned with the ways in which language use has changed  Sociolinguistics 147 over time; and a snapshot of a moment in time, usually contemporary, on the ‘syn- chronic’ axis. All the tools of linguistics may be deployed to focus in on particular features along these two dimensions, as we will outline in the rest of this chapter. Issues in sociolinguistics Sociolinguistics is a fieldwork-based discipline. Researchers collect examples of lan- guage usage in their naturally occurring environments and study them in relation to the findings of other sociolinguists’ research work. In this sense it is truly an example of applied linguistics: there is no introspection, nor intuitive conclusions, nor impres- sionistic evaluation involved. This means it is relatively easy for researchers new to the discipline to engage in genuine and valuable sociolinguistic research at an early stage in their study. Indeed, this sort of practical investigation would be the best way to develop your own thinking and knowledge of sociolinguistics. In order to demonstrate this fact, we introduce the key ideas in the field by illustration, using the sociolinguistic fieldwork data of Carmen Llamas. This research concentrates on the area of Teesside in the north-east of England, although the techniques Llamas uses and several of her findings are connected to many published sociolinguistic studies (Wol- fram and Schilling-Estes, 1998; Foulkes and Docherty, 1999; Kerswill, Llamas and Upton, 1999; Llamas, 2001, 2006, 2007a; Llamas and Watt, 2009). Categorizing the ways people speak Idiolect and sociolect Individuals speak in characteristic ways that might be peculiar to them in certain cir- cumstances: we call this pattern their ‘idiolect’. However, people often use language in ways that they share with many other people: most generally we can call these patterns ‘sociolects’. In part, the sociolects that individuals use help us to define them as a coherent social group. Sociolinguistics is mainly interested in the different forms of sociolect, in suggesting patterns and frameworks by which such sociolects seem to operate. It is a process of generalization away from the detail of specific data. In doing this, sociolinguistics does not deny the value of individual experience; indeed, the fact that social patterns are made explicit can be of immense value in understanding the place of individuals in society. Standard, non-standard and codification An example of the potential conflict that might result from these patterns can be seen in the tension – felt in almost all languages around the world – between the ‘standard’ form and ‘non-standard’ varieties. Standardization is a process that is apparent in almost all modern nations, in which one variety of a particular language is taken up (by government, the education system, newspapers and other media) and promoted as the ‘standard’ form. This often involves prescribing its use in the classroom and public examinations, reporting the workings of government in this form, printing national publications and any formal or prestigious material through its medium, and treating it 148 Llamas and Stockwell as the ‘correct’ and ‘proper’ form of the language (when, technically, there is no such thing). ‘Codification’ is a prominent feature of standard forms: grammar books and dictionaries are written promoting the form; texts of religious or cultural significance and canonical literature in the form are valued; and the variety is taught to children in schools (see Pennycook, 1994; Bex and Watts, 1999; Milroy and Milroy, 1999; Mug- glestone, 2003). Prestige, stigmatization and language loyalty By contrast, other non-standard forms of the language can be treated as ‘poor’ or ‘incorrect’ varieties: they are ‘stigmatized’. Standard forms receive ‘prestige’. It is easy to measure the relative prestige or stigma of a variety by asking the following questions:  Has the variety been ‘standardized’ and codified institutionally?  Is the variety spoken by a ‘living community’ of speakers?  Do the speakers have a sense of the long ‘history’ of their variety?  Do the speakers consider their variety to be independent of other forms and ‘autonomous’?  Do the speakers use the variety for all social functions and in all contexts or does it have a ‘reduced scope’ of usage?  Do the speakers consider their variety ‘pure’ or a ‘mixture’ of other forms?  Are there ‘unofficial’ rules of the variety, even where there is no codified grammar book; is there a sense of a ‘good’ and ‘bad’ form? (List adapted from Bell, 1976.) You will notice that these factors of prestige and stigmatization depend very much on speakers’ attitudes to their own variety. This is an important feature of socio- linguistic enquiry. People’s attitude to their own language often affects the form of that language. For example, stigmatized varieties of language often survive even under institutional pressure because groups have a ‘language loyalty’ that preserves the vari- eties in the face of the standardized form (see Garrett, Coupland and Williams, 2003). Dialect, accent and language planning A standardized variety is usually a regional ‘dialect’, which has been elevated in pres- tige and often loses its regional associations as a result. A dialect refers to the char- acteristic patterns of words and word-order (lexico-grammar) which are used by a group of speakers. The standard form of a language is an institutionally-valued dialect, which has been selected by historical accident or by deliberate ‘language planning’ by governments to be held up as the standard language. Dialect usually refers just to the form of the lexico-grammar of the variety as it could be written down, rather than its pattern of pronunciation. The latter is called ‘accent’. An accent can also be standardized and stigmatized. It is important to realize that accent and dialect are separate concepts. In principle, any dialect can be spoken in any accent, for example, the dialect known as Standard UK English can be heard in all of the regional accents of Britain. In practice, non-standard dialects tend to be spoken in  Sociolinguistics 149 specific local accents: it would be very strange (though possible) to hear a Liverpool dialect spoken in a New York accent, for instance. However, we often hear regional dialects spoken in foreign accents when they are being learned by non-native speakers. It is also important to realize that every form of spoken language is uttered as a dialect and in an accent. When people say they have no accent, they usually mean that they are speaking in a standardized and prestigious accent. Speech communities The way people speak often serves to define them as a group. We can talk of the ‘speech community’, which might correspond with the group as defined by other non- linguistic means: nationality, age range, gender, town or city population, political alle- giance and so on. As we will see in this chapter, the coherence generated by all these factors – including the linguistic factor – can operate as a self-serving reinforcement of all sorts of social values to do with local or community or ethnic identity. Language variants may also be maintained and reinforced, even against standardization pressure, in this way. Descriptive tools of language variation Any single piece of language is an integrated whole, but in order to investigate its dif- ferent aspects we must explore it in convenient categories. Traditionally, linguistics has categorized the different dimensions of language as a ‘rank scale’ from the smallest units of individual sounds or letters up to the largest scale of whole texts and dis- courses. Each of these levels often corresponds with a linguistic sub-discipline, as follows: Language element Linguistic sub-discipline discourse discourse analysis text text linguistics utterance pragmatics { sentence meaning → semantics clause & phrase structure → syntax word/lexeme lexicology morpheme morphology sound/phoneme phonology letter/grapheme graphology (For an overview of all these dimensions, see McGregor, 2009; Jackson and Stock- well, 2010; Mullany and Stockwell, 2010.) You will have noticed that some of the chapters in the first part of this book cover several of these sub-disciplines. Like second 150 Llamas and Stockwell language acquisition and psycholinguistics in this part of the book, the sub-discipline of sociolinguistics is not confined to one of these levels; instead, it investigates different levels from a sociolinguistic perspective. Although sociolinguistic variation occurs throughout the language system, socio- linguistic studies have focused on particular types of patterns, especially at the phonological level. Phonological variation is a useful level to study since it is easier to find an occurrence of a particular sound rather than a word, phrase or gram- matical structure; also, phonological variation is often below the level of awareness of speakers and so is less affected by self-conscious alteration. However, socio- linguistic exploration has also been undertaken at the grammatical, lexical, dis- coursal and whole-language levels. The linguistic variable The main tool in sociolinguistics has been the concept of the ‘linguistic variable’. This is any single feature of language which could be realized by different choices. In the word farm, for example, some people do not pronounce the /r/ and some do, and there are also variations in the ways the /r/ can be pronounced. This is a linguistic variable which is strongly determined by geographical location: non-/r/-pronouncers are likely to be from England, Wales, Aus- tralia, Massachusetts or the southern states of the USA. Furthermore, you could pronounce the /r/ as a sort of ‘tap’ against the back of the teeth (almost like a /d/), in which case you are likely to be from the Scottish Highlands or the west of Ireland. The linguistic variable feature could be a sound, or a word, or a phrase, or a pattern of discourse and so on. For example, common words for round bread products include the lexical variants: bun, roll, cob, bap, barm, fadgie, stotty, cake, batch, loaf and no doubt many others. You might not even recognize some of these, but their use is determined by the social factor of geographical location. Do you park your car, rank it or file it? Do you buy sugar in a bag, or a sack, or a poke? Do you call someone or phone them up or ring them or give them a phone or give them a bell or give them a buzz? All of these will vary depending on where you live, and who you are talking to. Phonological variation Although the linguistic variable can be from any level of the linguistic rank structure, it is variation in ‘accent’ that has provided the major focus of sociolinguistic studies so far. This is partly because observing and recording occurrences of individual sounds is very much easier than waiting around all day for a particular word, structure or dis- course pattern to appear, or setting up a complicated and artificial test situation. Pho- nological variables also have the advantage that they are usually below the level of conscious awareness, so the recorded data can be relied on to be naturalistic. People ordinarily talk of ‘broad’ or ‘strong’ accents and describe sounds as ‘precise’ or ‘clipped’ or a ‘drawl’. However, in order to be able to describe accents systematically and precisely, sociolinguists use the International Phonetic Alphabet (IPA). This is a system of special letters, each one of which corresponds with a very particular sound. The full IPA covers every speech sound it is possible to make with the human mouth and throat (see Ball and Rahilly, 1999; Collins and Mees, 2008). Table 9.1 lists a selection of some symbols which you might find useful in sociolinguistics. Table 9.1 Selected International Phonetic Alphabet (IPA) symbols Consonants p – pit tʃ – change b – bit dʒ – judge t – ten m – man d – den n – man k – cat η – sing g – get l – let f – fish r – ride, parrot (‘trilled r’) v – van ɾ – rubbish (Scots) (‘tapped r’) θ – thigh ɹ – farm (US) (‘approximant r’) ð – thy ʋ – ‘very’ as ‘vehwy’ s – set w – wet z – zen j – yet ʃ – ship ʔ – bu’er, ‘butter’ pronounced without the /t/’ (glottalstop) ʒ – leisure x – loch (Scots) h – hen Vowels (Monophthongs) (Diphthongs) ɪ – pit ɑɪ – bite, night ɛ – pet əɪ – night (Scots, Canadian) æ – pat ɛɪ – bait ɒ – pot (British accent) ɔɪ – boy ʌ – putt (British), color (US) əʊ – roe ʊ – put ɑʊ – house ə – patter (British) ʊə – sewer, poor (British) o – eau (French), low (N England) ɪə – ear (British) ɑ – calm (Scouse), farm (Teesside) ɛə – air (British) y – tu (French), school (Scouse) ø – peu (French), boat (Geordie) iː – bean ɜː – burn ɑː – barn ɔː – born uː – boon eː – bait (Northern England) 152 Llamas and Stockwell Grammatical variation Linguistic variables operating at a grammatical level have also been studied in socio- linguistics. For example, variations in the morphology of subject–verb agreement have been observed among the speech of British schoolchildren. The third person morpheme ‘-s’ (he goes, she knows) was used by some children for all verb agreements (I goes, I knows). It was noted that this non-standard pattern tended to be used with a greater fre- quency by boys than girls, and seemed to be a marker of group solidarity among the boys. Centrality in the social group and speech community is often marked by the frequent use of certain realizations of linguistic variables. A major feature of African–American vernacular English (AAVE) is the non-use of the verb ‘to be’ in some contexts: he a big man, you the teacher. This is known as ‘zero copula’, and is the grammatical form to use when the verb could be contracted in general American English or standard British Eng- lish: he’s a big man, you’re the teacher. By contrast, African–American vernacular has developed an invariant ‘be’ to signal habitual states: he be busy, she be running all day. A common grammatical variable that AAVE shares with many other non-standard grammars is the requirement for ‘negative concord’: that is, in a negated sentence, every element must be negated (Ain’t nobody going to help you, don’t nobody know me). This can be used for heavy emphasis (Ain’t no cat can’t get in no coop), where standard Englishes would need to use a few more phrases to achieve the same effect (There isn’t a single cat that can get into any coop at all) (see Labov, 1972; Kochman, 1981). Lexical variation Dialectal variation depends largely on different lexical items being used from region to region. Traditionally, ‘dialectologists’ were able to draw lines across maps in order to delineate the boundaries where different words or phrases were used. When making tea, you might stew, mash, brew or draw the tea in boiling water. Most local areas have specific lexical items that serve to identify their speakers: your nose is a neb in Yorkshire; a square is to Philadelphians what a block is to a New Yorker; an American resume is a British CV, which is South African biodata; South African robots are British traffic lights; American police batons are British truncheons which are Indian lathis and so on. Phrasal variations include the Irish and Scottish Is that you? when an English person would say Are you finished? and an American would say Are you done? or Are you through? Prepositional variation is very difficult to explain: why do Americans talk with and meet with when British people talk to and just meet? Something in back of the house in America is behind or at the back of it in Britain. There are dozens of others, usually consequences of historical divergence or interference from other languages. Discoursal variation Variability in discourse organization is a very fruitful area of investigation at the moment. Strategies of conversational structure can be observed and analysed, for example, and it is easy to see how politicians can be trained to exploit techniques for ‘keeping their turn’ (see Chapter 14, Speaking and pronunciation) and dominating the discussion. Alternatively, the different ways that men and women organize narratives or  Sociolinguistics 153 conduct conversations or arguments have been investigated to show up apparently dif- ferent objectives in speech. Aspects of politeness and social solidarity represent another dimension of discourse organization that can be explored (see Chapter 5, Pragmatics). Again, gender studies have led the way here, and insights into how politeness (and impoliteness) works have been generalized cross-culturally in comparative studies. The discoursal end of socio- linguistics is considered by some researchers to belong to pragmatics. Linguistic variation Lastly, the entire language can be treated as a variable. Bilingual or multilingual indi- viduals can often move from one language to another within a single utterance and sometimes even within a sentence. This is called ‘code-switching’, and the shift into another language can be used to indicate that a different ‘domain’ of experience is being signalled. Sometimes entire speech communities share two or more languages, as in Switzer- land (German, French, Italian) or Canada (French, English). Where there is a func- tional division between the languages’ usage, for example when one is used for formal or printed contexts and the other just in speech, then a situation of ‘diglossia’ is said to exist. One variety becomes the H (as in High German) and the other the L (Low German) variety. For example, classical Arabic, the language of the Koran, is the H variety that can be read by all Arabic speakers, but in different Arab countries a range of different L varieties of Arabic is spoken. Sociolinguistics explores aspects of such situations, as well as deliberate attempts by governments and authorities to engage in language planning: the promotion and stan- dardization of one variety of language, and attempted interventions in linguistic usage (such as Noah Webster’s dictionary with its new spellings of ‘American English’ words, or prohibitions by the Academie Française of Anglicisms such as le weekend or le hot- dog in French). Lastly, sociolinguists explore the birth and death of languages, for example in the development of ‘pidgin’ languages. These are new languages, often based on two or more languages in contact, with their own systematic grammatical rules. When some pidgins become the first languages of a new generation, they are called ‘creoles’ (such as South African Afrikaans, Jamaican Patwa, West African Krio, Louisiana Creole and many others). Creolists have provided insights into the pro- cesses of development of all languages, by investigating new and emerging creoles (see Holm, 1988, 1989; Kouwenberg and Singler, 2005; Mufwene, 2001; Romaine, 1988; Sebba, 1997). Social factors that correlate with language variation In the section above, it was very difficult for us to talk about linguistic variables with- out mentioning the social factors with which they may correlate. This is the whole point of sociolinguistics. In investigation, a linguistic variable is set against the social variable in order to work out the influence of that social aspect on language. A range of social variables has been focused upon in sociolinguistic studies (see Llamas, Mullany and Stockwell, 2007; Milroy and Milroy, 1993). 154 Llamas and Stockwell Geographical and social mobility Dialects within a language are often localized geographically. We can speak of ‘dialect chains’ where the shift from one dialect to the next is not sudden between one town or county or state and the next. Instead, dialects merge and overlap across distances. Even at national boundaries, speakers on either side of the border can sometimes understand each others’ dialects (such as neighbouring Dutch and Germans) better than speakers within their own ‘language’ community (northern Germans and Bavarian Germans, for example). If dialect chains complicate the dialect map, towns and cities complicate matters further. The migration of people into urban areas disrupted neat dialect divisions, and the study of ‘urban dialectology’ was only achieved by the realization that there is social stratification in urban areas on the basis of class. Increasing geographical mobility has been matched over the last century in the Western world by increasing social mobility. The self-con- sciousness that this brings can be observed in people of certain social groups aiming for a more prestigious form of language than they would naturally use, for example, ‘overdoing’ an upper-middle class accent in formal situations. This is called ‘hypercorrection’. The counterpart of hypercorrection is the phenomenon observed when some people use stigmatized forms of language (as a sort of ‘streetwise’ accent signal): this is known as ‘covert prestige’. Factors such as these are major influences on language loyalty and language change. Gender and power The influence of gender and asymmetries in power relations have been a major aspect of sociolinguistic discussion in recent years. The notion of a ‘genderlect’ has been pro- posed to account for some of the apparently systematic differences in the ways men and women use language. These differences can be observed across the whole range of lin- guistic variables, from plans of narrative and discourse organization, to the different accents that men and women have even from the same area (see Coates and Cameron, 1986; Cameron, 1995; Crawford, 1995; Mills, 1995; Holmes and Meyerhoff, 2003; Coates, 2004). Age Older people and younger people use language differently. When corresponding fea- tures of these speakers are compared, such differences can reveal evidence of changes in the language over time. In other words, the ‘snapshot’ of current usage across the age ranges can suggest historical language changes. This is the ‘apparent time hypothesis’; it gives us the ability to observe potential change in progress, which was not thought possible in the past (see Llamas, 2007b). Audience Taking into account the audience and reception of language use provides insights into the ways speakers behave. Most conversations have a ‘recipient design’, that is, speakers plan their utterances with the addressee in mind. This factor often results in speakers adjusting their accent, style or language towards their addressees. This phenomenon is  Sociolinguistics 155 called ‘accommodation’ and it seems that such convergence of accents is an important cause of language change over time (see Auer, 2007). Identity This is an important social factor. Not only do linguistic patterns signal social and indi- vidual identity, but people’s conscious awareness of their personal, ethnic, geographical, political and family identities is often a factor in their language use. Allegiance and membership of different social groups can be expressed by language patterns, and some- times those groups are even defined by these patterns, whether this is a language or style or jargon (see Eckert, 2000; Dyer, 2007; Llamas and Watt, 2009; Mullany, 2007). Social network relations It has been recognized that the relative strength of relations between individuals within a social group (their ‘social network’) is also important in understanding how linguistic features are maintained, reinforced and spread. Whether individuals have strong or weak ties to the group can be used as a measure of their sociolinguistic influence (see Milroy, 1987; Milroy and Milroy, 1999). Working with sociolinguistic data Collecting and analysing sociolinguistic data When collecting data, the fieldworker must be aware of a range of issues involved in ‘sampling’ and the ‘representativeness’ of the population surveyed. A variety of tech- niques have been developed by sociolinguists to gain access to the least monitored forms of speech, below the level of common self-awareness. Among the ‘experimental’ forms of elicitation that can be used are interviews, ques- tionnaires (spoken or written), ‘thinking-out-loud protocols/think-aloud protocols’ (TOL/ TAP) given with a passage to read, role-play and storytelling. Linguists have also investi- gated speech styles by use of a series of elicitation techniques that have increasing degrees of informant self-awareness, for example, starting with an informal conversation, then giving a reading passage, then a list of words to read, and finally a list of potential minimal pairs (such as moon/moan, which/witch or cot/caught). Another example of an innovative technique was used by Llamas in her fieldwork: ‘a sense relation network’ sheet (also known as a ‘semantic map’) intended to elicit local speech variants. It is a fact of sociolinguistic research that if people are aware they are being observed, they often alter their linguistic behaviour. This is the ‘observer’s paradox’, and several of the methods above were developed in order to minimize its impact on the data collected. The ‘ethical’ consequences of data collection must also be con- sidered, in relation to the informants’ rights of privacy. Interpreting sociolinguistic data Now that we have introduced some of the key concepts involved in sociolinguistics and we have considered factors to bear in mind when collecting sociolinguistic data, we 156 Llamas and Stockwell must think about how we interpret the data we collect. As well as discovering variation, we must attempt to explain what motivates this variation. Some questions we must think about include:  Why does language variation exist (particularly variation between speakers from the same speech community)?  What function does the variation serve?  How do languages change?  What processes are involved?  Does the data we collect from one speech community have wider implications? In this section we shall use aspects of Llamas’s research on Teesside English to con- sider what linguistic data can tell us about the nature and function of language varia- tion and change. Before looking at data from the Teesside study, however, we shall consider some models and frameworks we work within when interpreting language variation and change. This will allow us insight into decisions made in the design of the Teesside study and the research questions it addresses. Models and frameworks The axiom underlying our initial definition of sociolinguistics is that language is vari- able at all times. Variation means there is the potential for change, and the causes and effects of language change are, therefore, central concerns of sociolinguistics. In seeking the motivation for language change, we must consider whether the changes are internal or external to the linguistic system. Internal changes are ‘system-based’, brought about by pressures internal to the linguistic system. For example, vowel changes affecting a number of northern cities in the USA are often explained from the perspective of a ‘chain shift model’. In this framework, changes in vowel sounds are co-ordinated, that is, movement of one vowel triggers movement of another and another and so on down the chain. Within sociolinguistics, external changes are ‘speaker-based’, brought about by speakers adopting forms from other varieties. The Teesside study focuses on varia- tion in the realization of certain consonants and considers the variation to be speaker- based. The motivation for the variation is thus seen as social and external to the lin- guistic system. Unprecedented changes have been witnessed in spoken British English in recent years, most of which appear to be best accounted for by factors external to the lin- guistic system. A ‘dialect levelling model’ of change has been used to account for data in a number of studies. ‘Dialect’ or ‘accent levelling’ involves the eradication of locally marked forms in a variety. Large-scale homogenization appears to be taking place in spoken British English: differences between accents are becoming less marked. A ‘gravity model’ of ‘diffusion’, which involves the spreading of variants from an identi- fiable local base into other geographical localities, also appears to be underway. Many of the spreading features in British English are thought to be moving northwards from a south-eastern epicentre. Forms associated with London English are now found in urban centres far from the capital. Both levelling and diffusion come about through the ‘dialect contact’ caused by geographical and social mobility. As people increasingly travel and move across society,  Sociolinguistics 157 speakers often experience considerable face-to-face contact with speakers of other varieties. In these contact situations, speakers tend to avoid very locally marked forms of speech (this is called ‘accommodation’, where speakers move towards their inter- locutor’s speech patterns). If this happens sufficiently frequently and in sufficiently large numbers, the accommodation can become permanent. Contact-based changes have often been thought to be changes towards the standard variety. However, non-standard varieties are exercising more and more influence in British English and many of the current changes in progress involve the spread of non-standard forms. Let us look at some evidence from the Teesside study to see whether our linguistic data can be interpreted by the models of change we have been considering. The Teesside study The study set out to investigate whether localized forms were coming under pressure from other vernacular forms spreading from outside the area. A previously unre- searched urban variety of British English was chosen as the locality for the research: Middlesbrough, the major urban centre of the conurbation around the River Tees, lies some 260 miles north of London, and offered a good case study situation. Llamas wanted to discover whether local forms were being eradicated and whether spreading vernacular forms had made inroads into Middlesbrough English (MbE). Combined with analysing variation within MbE, evidence for linguistic change in progress was sought in the study. For this reason, the two social variables of age and gender were included in the design of the fieldwork sample. Data were taken from a sample of 32 speakers from Middlesbrough who formed a socially homogeneous group, all being ‘working-class’ by their own self-assessment. In order to detect potential lin- guistic changes in progress using the ‘apparent time hypothesis’, four age groups of speakers were included in the sample (Table 9.2). Llamas conducted interviews with informants in self-selected pairs, using a new method of data elicitation. The method was designed to elicit data which are ana- lysable on five levels of the rank scale we looked at earlier: phonology, morphology, syntax, lexicology and discoursal variation (although only the first four were ana- lysed in the study). The principal research tool used in the interview was a Sense Relation Network sheet, where subjects were given prompts, such as tired, throw away or tell to be quiet, and then asked to write in alternative words or phrases from their own vernacular. So, let us look at some data from the study to see whether we can detect any sys- tematic variation in the sample or any evidence for possible change in progress in MbE. Table 9.2 Design of the fieldwork sample Old (60–80) Middle-aged (32–45) Young adult (19–22) Adolescent (16–17) Male Female Male Female Male Female Male Female 4 4 4 4 4 4 4 4 158 Llamas and Stockwell One of the linguistic variables included in the study was intervocalic /r/ (as in carry, area, a real and to reach). Three variants of /r/ were analysed in the data:  The alveolar tap [ɾ].  The alveolar approximant [ɹ].  The labio-dental approximant [ʋ]. Note: Whereas phonemes are represented in slashes, for example /r/, the various slightly different ways of pronouncing a phoneme are represented by square brackets, for example [ɾ], these are called ‘allophones’ (see Celce-Murcia, Brinton and Goodwin, 1996, and Collins and Mees, 2008, for a description of phonemes and allophones and how they are produced in the vocal tract).  [ɾ] This tap may be considered the ‘localized variant’ which is found in northern England and Scotland.  [ɹ] This alveolar approximant is the non-localized, or ‘standard variant’.  [ʋ] This labio-dental is the ‘spreading variant’ which is currently spreading rapidly from the south of England. Figure 9.1 reveals whether use of the localized variant, the standard variant and the spreading variant can be correlated with any of the social groups of speakers. What Figure 9.1 shows is that there is a great deal of variation in the use of variants of /r/ among the speakers of the sample. This variation is both gender-correlated and age-correlated. Figure 9.1 Distribution of variants of /r/ in Middlesbrough English  Sociolinguistics 159 If we consider the age variation first, the data suggest that use of the localized var- iant [ɾ] is in steady and dramatic decline (it is used by the old speakers, but almost rejected categorically by the young speakers). The age-correlated variation also sug- gests that [ʋ] is a new variant which has appeared in MbE very recently (it is used to a considerable extent by the young speakers, but not found at all in the speech of the old). What we have, then, is evidence suggesting change in progress in MbE. This change appears to involve the processes of both levelling and diffusion. There also seems to be a gender difference between the initiators of these processes of change. The findings indicate that the females lead in the levelling out of variants, with males following (note the much lower female use of [ɾ]). Males, on the other hand, lead in the diffusion of new variants into the vernacular, with females following (note the higher use of [ʋ] among young males). It seems, then, from the data for /r/ that MbE is indeed undergoing a process of levelling and features which are spreading from the south-east of England are appear- ing in the speech of the young in Middlesbrough. Does this mean that MbE is becoming like accents of the south-east of England? Let’s look at another variable. Intervocalic, word-medial /p/ was also taken as a variable in the study (as in ‘paper’). Three allophones of /p/ were under investigation:  The standard variant is the released bilabial stop [p].  Another possible variant is the glottal stop [ʔ].  The variant local to the north-east of England is a glottalized [ʔp] (this represents a simultaneous glottal stop and ‘p’ sound). Given the dialect levelling in evidence in the variable /r/, we may expect the same to be true of /p/ with a marked decline revealed in the use of [ʔp]. Let us see. Figure 9.2 Distribution of variants of /p/ in Middlesbrough English 160 Llamas and Stockwell The most immediately striking thing we see in Figure 9.2 is the marked gender difference. The women show a clear preference for the standard variant [p], whereas the men favour the localized [ʔp]. This type of gender-correlated variation has been found repeatedly in socio- linguistic studies. If we look closely at the data, however, we notice that the young women are acting quite differently from the old and middle-aged women. The young women demon- strate a much higher use of the localized north-eastern [ʔp]. Such is the increase in usage that [ʔp] is the preferred variant of the adolescent women compared with a 4.6 per cent use among the old female speakers. Far from being levelled out then, use of the localized variant of /p/ appears to be on the increase. Also, an increase is revealed in use of the glottal stop, in parti- cular among the young female speakers. Again, we have evidence which suggests change in progress in MbE as well as the existence of sharply differentiated genderlects. It is clear, then, from looking at just two linguistic variables and co-varying them with two social variables, that socially meaningful language variation can be detected, and from the evidence of variation we can infer patterns of change. Evidence from /r/ and /p/ both suggest that change is in progress in MbE. In both variables we also see that men speak differently from women of the same speech community, indeed, in many cases of the same family. The variation in language is clearly not random or free. Rather, it appears to be systematic and to be constrained by social factors. Although the groups we are working with are made up of individuals (a fact we should not forget), the individual speakers appear to systematically prefer or disprefer variants that are available to them depending on whether they are male or female, young or old. In this way speakers realize their sociolinguistic identity and are able to project the linguistic iden- tity they choose to the outside world. Gaining insight into the motivation for these choices is also part of our job as sociolinguists. The different variants must carry symbolic meaning to the speakers whether or not the speakers would be able to explain what that meaning is. By analysing other factors – how mobile the speakers are, how they evaluate the variants under consideration – we may find answers to some of the questions posed by our findings:  Why are some variants adopted from other varieties and others not?  Why are some variants in decline and others increasing?  Why are some variants preferred by female speakers and others by males? The data we have looked at have revealed another important fact to us. Although the increasing variants we have seen, [ʋ] and [ʔp], are different – one is new to MbE and the other is local to MbE – one thing they have in common is that they are both non-standard forms. The changes in progress that are suggested by the evident variation in the data, then, do not represent a movement of MbE towards the standard variety. This is true of many other localities and many other variables. The ‘covert prestige’ carried by non-standard forms seems to be exercising more and more influence on language variation and change. This suggests we should re-evaluate the influence of the standard variety, for example Received Pronunciation (RP) in British English, and to question its status as a model to be imitated in language teaching. Applications of sociolinguistics Many sociolinguistic studies have a practical application as their main objective. Sociolinguistics has informed the thinking of government policy on education and  Sociolinguistics 161 language planning across the world, with insights from the field finding their way directly into teacher-training courses and educational programmes, especially in the UK and USA. Teachers who are aware of the sociolinguistic context have insights at their disposal which can make them better teachers. For example, what was once regarded as ‘bad’ grammar can be seen as a systematic non-standard dialect, and corrective teaching can be replaced by an awareness of multidialectalism. This can give students a greater repertoire in their performance, including access to the prestigious standard forms, and a greater confidence in their own language abilities. It encourages us to recognize diversity as rich- ness. Lippi-Green (1997), for example, contains a wealth of information on how language prejudice and ideological planning have operated in the USA. There are many other uses of sociolinguistics. Film actors imitating accents will have been trained using insights from sociolinguistics. Criminals have been caught by pin- pointing their accent origins. Politicians, advertisers and assertiveness trainers all learn discourse patterns that convey their message most effectively. In addition, socio- linguistic studies have contributed greatly to our understanding of how languages change. For example, Labov (1994, 2001) and Milroy (1992) demonstrate a socio- linguistic view of the historical development of English. This not only helps us to ‘read’ the past but also offers us guidance on the likely social implications in the future. Finally, the methods developed in sociolinguistics have led the way in the con- sideration of research ethics and in the use of naturalistic data in linguistic study. Sociolinguistics reveals the complexity of context when language is studied in its real, applied setting, and it also suggests ways of understanding this context and the richness of language uses. Hands-on activity The passage below is a humorous attempt to imitate the spoken vernacular of Mid- dlesbrough in written form (and also pokes some gentle fun at sociolinguists). We have chosen this passage as it is likely to be unfamiliar to most of the readers of this book. However, you do not need to understand the passage at first to be able to use it as sociolinguistic data. First, draw a large table divided as follows: Middlesbrough English Standard English Phonological examples (Any novel spellings that seem to be used to represent the accent) Lexical examples (Any words you do not recognize, or which seem to be used in an unusual way) Grammatical examples (Including strange idioms, as well as unu- sual phrases and syntactic ordering that you find odd) Discoursal examples (Anything which seems to be trying to capture spoken discourse)