Sociolinguistics 2 PDF

6 Regional and social dialects I n the first section of this book, the focus was on language variation in multilingual communities. In this section, the focus moves to language variation in monolingual communities. People often use a language to signal their membership of particu...

6 Regional and social dialects I n the first section of this book, the focus was on language variation in multilingual communities. In this section, the focus moves to language variation in monolingual communities. People often use a language to signal their membership of particular groups and to construct different aspects of their social identity. Social status, gender, age, ethnicity and the kinds of social networks that people belong to turn out to be important dimensions of identity in many communities. I will illustrate the way people use language to signal and enact such affiliations in this second section of this book. Example 1 Telephone rings. Pat: Hello. Caller: Hello, is Mark there? Pat: Yes. Just hold on a minute. Pat (to Mark): There’s a rather well-educated young lady from Scotland on the phone for you. When you answer the telephone, you can often make some pretty accurate guesses about various characteristics of the speaker. Pat was able to deduce quite a lot about Mark’s caller, even though the caller had said nothing explicitly about herself. Most listeners can identify that the caller is a child without any problem. When the caller is an adult, it is usually easy to tell whether a speaker is female or male. If the person has a distinctive regional accent, then their regional origins will be evident even from a short utterance. And it may also be possible to make a reasonable guess about the person’s socio-economic or educational background, as Pat did. No two people speak exactly the same. There are infinite sources of variation in speech. A sound spectrograph, a machine which represents the sound waves of speech in visual form, shows that even a single vowel may be pronounced in hundreds of minutely different ways, most of which listeners do not even register. Some features of speech, however, are shared by groups, and become important because they differentiate one group from another. Just as different languages often serve a unifying and separating function for their speakers, so do speech characteristics within languages. The pronunciation, grammar and vocabulary of Scottish speakers of English is in some respects quite distinct from that of people from England, for example. Though there is variation within Scotland, there are also 131 An introduction to sociolinguistics some features which perform an overall unifying function. The letter r in words like girl and star is pronounced in a number of English-speaking areas, and Scotland is certainly one of them. And a Scot is far more likely to say I’ll not do it than I won’t do it. Similarly the pronunciation of bath with the same vowel as in sat distinguishes a speaker from the north of England from a southerner. And while many speakers of English use the same vowel in the three words bag, map and bad, workers in Belfast pronounce them in ways that sound like [beg], [ma::rp] and [bod] to English people. Speech provides social information too. Dropping the initial [h] in words like house and heaven often indicates a lower socio-economic background in English. And so does the use of grammatical patterns such as they don’t know nothing them kids or I done it last week. We signal our group affiliations and our social identities by the speech forms we use. Regional variation International varieties Example 2 A British visitor to New Zealand decided that while he was in Auckland he would look up an old friend from his war days. He found the address, walked up the path and knocked on the door. ‘Gidday,’ said the young man who opened the door. ‘What can I do for you?’ ‘I’ve called to see me old mate Don Stone,’ said the visitor. ‘Oh he’s dead now mate,’ said the young man. The visitor was about to express condolences when he was thumped on the back by Don Stone himself. The young man had said, ‘Here’s dad now mate’, as his father came in the gate. There are many such stories – some no doubt apocryphal – of mistakes based on regional accent differences. To British ears, a New Zealander’s dad sounds like an English person’s dead, bad sounds like bed and six sounds like sucks. Americans and Australians, as well as New Zealanders, tell of British visitors who were given pens instead of pins and pans instead of pens. On the other hand, an American’s god sounds like an English person’s guard, and an American’s ladder is pronounced identically with latter. Wellington sux Auckland nil Graffiti on a wall in Wellington There are vocabulary differences in the varieties spoken in different regions too. Australians talk of sole parents, for example, while people in England call them single parents, and New Zealanders call them solo parents. South Africans use the term robot for British traffic-light. British wellies (Wellington boots) are New Zealand gummies ( gumboots), while the word togs refers to very different types of clothes in different places. In New Zealand, togs are what you swim in. In Britain you might wear them to a formal dinner. 132  Chapter 6 Regional and social dialects Exercise 1 You may like to check out the extent of US vs British influence on vocabulary in your region. The following questions provide a simple way of measuring this. Ask ten of your friends to answer them and work out how many US items vs how many British items they choose. You should allow for the fact that some may use both. If you are not sure which is the British item and which is used in the USA, check in a big reference dictionary such as Webster’s Third New International Dictionary or the big Oxford English Dictionary. (a) When you go window-shopping do you walk on the pavement or the sidewalk? (b) Do you put your shopping in the car’s trunk or in the boot? (c) When the car’s engine needs oil do you open the bonnet or the hood? (d) Do you fill up the car with gas or with petrol? (e) When it is cold do you put on a jersey or a sweater? (f) When the baby is wet does it need a dry diaper or nappy? (g) Do you get to the top of the building in an elevator or a lift? (h) When the children are hungry do you open a can or a tin of beans? (i) When you go on holiday do you take luggage or baggage? (j) When you’ve made an error do you remove it with an eraser or a rubber? Example 3 (a) Do you have a match? (b) Have you got a cigarette? (c) She has gotten used to the noise. (d) She’s got used to the noise. (e) He dove in, head first. (f ) He dived in head first. (g) Did you eat yet? (h) Have you eaten yet? Pronunciation and vocabulary differences are probably the differences people are most aware of between different dialects of English, but there are grammatical differences too. Can you distinguish the preferred US usages from the traditional British usages in the sentences in example 3? Speakers of US English tend to prefer do you have, though this can now also be heard in Britain alongside the traditional British English have you got. Americans say gotten where people in England use got. Many Americans use dove while most British English speakers prefer dived. Americans ask did you eat? while the English ask have you eaten? Are the US or the British usages predominant where you live? In New Zealand, where US forms are usually regarded as more innovative, younger New Zealanders say dove, while older New Zealanders use dived. The differences that English speakers throughout the world notice when they meet English speakers from other nations are similar to those noted by speakers of other languages too. Spanish and French, for example, are languages which are extensively used in a variety of countries besides Spain and France. Speakers of Spanish can hear differences of pronunciation, vocabulary and grammar in the varieties of Spanish spoken in Mexico, Spain, Argentina and 133 An introduction to sociolinguistics Paraguay, for example. Native speakers of French can distinguish the French used in Montreal from Parisian and Haitian French. There are differences in the vocabulary of different varieties. So, for example, a Parisian’s travail (‘work’) is a djobe in Montreal. The word for ‘beggar’ is mendiant in France but quêteux in Quebec. And Canadians tend to use aller voir un film, while Parisians prefer aller au cinéma. Even grammatical gender assignment differs in the two varieties. Appétit (‘appetite’) and midi (‘midday’), for instance, are feminine in Canada, but masculine in France, while the opposite is true for automobile and oreille (‘ear’). Clearly Canadian French and Parisian French are different dialects. Sometimes the differences between dialects are a matter of the frequencies with which particular features occur, rather than completely different ways of saying things. People in Montreal, for example, do not always pronounce the l in phrases like il pleut and il fait. Parisians omit the l too – but less often. If you learned French in school you probably struggled to learn which verbs used avoir and which used être in marking the perfect aspect. Getting control of these patterns generally causes all kinds of headaches. It would probably have caused you even more pain if you had realised that the patterns for using avoir and être are different in Montreal and Paris. Exercise 2 How do you pronounce batter? How many different pronunciations of this word have you noticed? Use any method you like to represent the different pronunciations. Answer at end of chapter Intra-national or intra-continental variation Example 4 Rob: This wheel’s completely disjaskit. Alan: I might could get it changed. Rob: You couldn’t do nothing of the sort. It needs dumped. This conversation between two Geordies (people from Tyneside in England) is likely to per- plex many English speakers. The double modal might could is typical Geordie, though it is also heard in some parts of the southern USA. The expression needs dumped is also typical Tyneside, though also used in Scotland, as is the vocabulary item disjasket, meaning ‘worn out’ or ‘completely ruined’. The way English is pronounced is also quite distinctive in Tyneside, and perhaps especially the intonation patterns. Because they like the speech heard in television programmes such as Auf Wiedersehen Pet, Byker Grove and Joe Maddison’s War, some people can imitate the tune of Geordie speech – if nothing else. We are dealing here not just with different accents but with dialect differences within a country, since the distinguishing forms involve grammatical usages and lexical items as well as pronunciation. Regional variation takes time to develop. British and US English, for instance, provide much more evidence of regional variation than New Zealand or Australian English. Dialectologists can distinguish regional varieties for almost every English county, e.g. Yorkshire, Lancashire, Northumberland, Somerset, Cornwall and so on, and for many towns too. Some British 134  Chapter 6 Regional and social dialects dialects, such as Scouse (heard in Liverpool), Cockney and Geordie, even have distinct names showing how significant they are in distinguishing groups from one another. Within the London area, the Cockney dialect is quite distinctive with its glottal stop [c] instead of [t] between vowels in words like bitter and butter, and its rhyming slang: e.g. apples and pears for ‘stairs’, lean and lurch for ‘church’, the undoubtedly sexist trouble and strife for ‘wife’ and the more ambiguous cows and kisses for ‘the missus’. In the USA, too, dialectologists can identify distinguishing features of the speech of people from different regions. Northern, Midland and Southern are the main divisions, and within those three areas a number of further divisions can be made. Different towns and even parts of towns can be distinguished. Within the Midland area, for example, the Eastern States can be distinguished; and within those the Boston dialect is different from that of New York City; and within New York City, Brooklynese is quite distinctive. The Linguistic Atlas Projects (http://us.english.uga.edu/) provide a rich source of information on the features of pronunciation, grammar and vocabulary which distinguish different US dialects. In the rural Appalachians, one can hear pronunciations such as acrosst and clifft, as well as verbs with a-prefixes, such as a-fishin’ and a-comin’. Words for dragonfly in the Eastern States include darning needle, mosquito hawk, spindle, snake feeder, snake doctor and snake waiter, but of these only darning needle is used in New York. From darning needle, however, New York has developed two new variants dining needle and diamond needle. (It becomes difficult at this point to remember that these are all names for an insect not a sewing implement!) In areas where English has been introduced more recently, such as Australia and New Zealand, there seems to be less regional variation – though there is evidence of social variation. The high level of intra-national communication, together with the relatively small populations, may have inhibited the development of marked regional differences in these countries. In New Zealand, for instance, there are greater differences among the Maori dialects than within English, reflecting the longer period of settlement and more restricted means of communication between people from different Maori tribes before European settlers arrived. Maori pronunci- ation of words written with an initial wh, for example, differs from one place to another. The Maori word for ‘fish’ is ika in most areas but ngohi in the far North, and kirikiri refers to ‘gravel’ in the west but ‘sand’ in the east of New Zealand. There are many more such differences. Exercise 3 Can you guess what the following words and phrases mean? They are all words collected from regional dialects of British English. A good British English dialect dictionary will provide information about their meanings and where they are used. (a) snowblossom (b) time for our snap (c) mask the tea (d) the place was all frousted (e) clinker bells (f) a great mawther (g) I’m really stalled (h) a bairn (i) an effet (j) I’ll fill up your piggy, it’s time for bed Answers at end of chapter 135 An introduction to sociolinguistics Figure 6.1 Words for splinter in English dialects Source: Trudgill (1994: 21). Figure 6.1 is a map of England showing where different dialect words are used for the standard English word splinter. The boundary lines are called isoglosses. This is just one word out of thousands of linguistic features which vary in different dialects, and which were documented by Harold Orton’s comprehensive Survey of English Dialects in the 1950s. When all the informa- tion on linguistic regional variation is gathered together on a map, with isoglosses drawn between areas where different vocabulary, or grammatical usages or pronunciations occur, the result looks something like a spider’s web. Some of the web’s lines are thicker than others because a number of boundaries between features coincide. But there is also a great deal of overlap between areas. The line between an area where people use [a] rather than [a:] in a word like path, for example, does not coincide with the line which separates areas using have you any sugar? rather than have you got any sugar? Areas which use the word elevenses rather than snap or snack do not all use different words for brew or snowflake or manure or 136  Chapter 6 Regional and social dialects splinter. The same vocabulary may be used throughout an area where contrasts in the pro- nunciation of words are quite dramatic. In other words, defining linguistic areas is not at all straightforward – a point which is illustrated very clearly in example 5 below. Exercise 4 Where there are differences between regions, it is interesting to discover the local names for particular objects. There are often regional differences in the words used for standard English scarecrow, stream and cowpat, for instance. When asking people what they call these items, you should phrase your question so as to avoid using the word you are interested in. To exemplify, I have provided four questions aimed at eliciting labels for four more objects which often vary regionally: (a) What do you call a small round sweet cake with a hole in the middle? (b) What do you call the vehicles people push babies round in? (c) What do you call an item of clothing worn to protect clothing especially while cooking? (d) What do you call the shoes people wear for tennis or running? Collect information from a range of people on what they call these objects and where possible include older people who were born outside your area. Cross-continental variation: dialect chains Example 5 Miriam learnt French and Italian at university and was a fluent speaker of both. As part of her course she was required to study for three months in Paris and three months in Rome. Her time in Paris went well and she decided to take a holiday on her way to Rome, travelling across France to Italy. She was keen to hear the varieties of French and Italian spoken in provincial towns. She stayed in cheap pensions (French ‘bed-and-breakfast’ places), and she made a special effort to talk to the local people rather than tourists. Her Parisian accent was admired and she could understand the French of Dijon and Lyon. But as she moved further from Paris she found the French more difficult to follow. Near the border between France and Italy, in the town of Chambéry, she could not be sure what she was hearing. Was it Italian French or French Italian? Whatever it was, it was difficult for her to understand, though she had no trouble making herself understood. Most people thought she spoke beautifully – especially for a foreigner! In Italy she found that the Italian spoken in Turin and Milan was very different from the Italian she had learned. As she approached Rome, however, she gradually began to comprehend more of what she heard. And finally in Rome she found some kind of match between the way she spoke and the way the Italians around her spoke. Though a map suggests the languages of Europe or India are tidily compartmentalised, in reality they ‘blend’ into one another. The varieties of French spoken in the border towns and villages of Italy, Spain and Switzerland have more in common with the language of 137 An introduction to sociolinguistics the next village than the language of Paris. From one village and town to the next there is a chain or continuum. Dialect chains are very common across the whole of Europe. One chain links all the dialects of German, Dutch and Flemish from Switzerland through Austria and Germany, to the Netherlands and Belgium, and there is another which links dialects of Portuguese, Spanish, Catalan, French and Italian. A Scandinavian chain links dialects of Norwegian, Swedish and Danish, so that Swedes and Norwegians in adjacent areas can communicate more easily than fellow-Swedes from southern and northern Sweden. The same kind of dialect chains are found throughout India and China. They illustrate very clearly the arbitrariness of the distinction between ‘language’ and ‘dialect’. It is easy to see that if we try to define what counts as German vs Dutch or Swedish vs Norwegian or Italian vs French using only linguistic features, the task will be fraught with problems. Where should we draw the boundaries between one dialect and the next, or one language and the next? The linguistic features overlap, and usage in one area merges into the next. Intelligibility is no help either. Most Norwegians claim they can understand Swedish, for instance, although two distinct languages are involved, while Chinese who speak only Cantonese cannot understand those who speak Mandarin, despite the fact that both are described as dialects of the Chinese language. Example 6 Ming is an elderly woman who lives with her son in a rural village near the town of Yingde in Guangdong Province in southern China. The family grows vegetables for the local market. Ming speaks only her provincial dialect of Chinese, Cantonese. Last summer, Gong, an official from Beijing in the north, visited her village to check on the level of rice and ginger production. Gong also spoke Chinese, but his dialect was Mandarin or putonghua. Ming could not understand a single word Gong said. Languages are not purely linguistic entities. They serve social functions. In order to define a language, it is important to look to its social and political functions, as well as its linguistic features. So a language can be thought of as a collection of dialects that are usually linguist- ically similar, used by different social groups who choose to say that they are speakers of one language which functions to unite and represent them to other groups. This definition is a sociolinguistic rather than a linguistic one: it includes all the linguistically very different Chinese dialects, which the Chinese define as one language, while separating the languages of Scandinavia which are linguistically very similar, but politically quite distinct varieties. Exercise 5 (a) Define the difference between a regional accent and a regional dialect. (b) What do you think is the difference between a regional and a social dialect? Answers at end of chapter 138  Chapter 6 Regional and social dialects Social variation RP: a social accent Example 7 Diana: Have you heard – Jonathan’s engaged to that northern girl – from Cumbria! Reg: She may be northern but I assure you she is very acceptable. Her father is a lord, and a rich one at that! She has had the best education money can buy. Those traces of northern accent are fashionable these days my dear! In earlier centuries, you could tell where an English lord or lady came from by their regional form of English. But by the early twentieth century, a person who spoke with a regional accent in England was most unlikely to belong to the upper class. Upper-class people had an upper-class education, and that generally meant a public (i.e. private!) school where they learned to speak RP. RP stands not for ‘Real Posh’ (as suggested to me by a young friend), but rather for Received Pronunciation – the accent of the best educated and most prestigious members of English society. It is claimed that the label derives from the accent which was ‘received’ at the royal court, and it is sometimes identified with ‘the Queen’s English’, although the accent used by Queen Elizabeth II, as portrayed so brilliantly by Helen Mirren in the movie The Queen, is a rather old-fashioned variety of RP. RP was promoted by the BBC for decades. It is essentially a social accent not a regional one. Indeed, it conceals a speaker’s regional origins. This is nicely illustrated in figure 6.2, the accent triangle. As the triangle suggests, most linguistic variation will be found at the lowest socio- economic level where regional differences abound. Further up the social ladder the amount of observable variation reduces till one reaches the pinnacle of RP – an accent used by less than 5 per cent of the British population. So a linguist travelling round Britain may collect over a dozen different pronunciations of the word grass from the working-class people she meets in different regions. She will hear very much less variation from the lower-middle-class Figure 6.2 Social and regional accent variation Source: From Peter Trudgill (1983) Sociolinguistics. Penguin: London. Page 42. Copyright © Peter Trudgill, 1974, 1983, 1995, 2000. Reproduced by permission of Penguin Books Ltd. 139 An introduction to sociolinguistics and middle-class people. And, at least until recently, the upper classes would pronounce the word as [gra:s] wherever they came from in England. Things are changing, however, as the exchange in example 7 suggests. Figure 6.2 captured the distribution of accents in England until recently. Today a more accurate diagram might have a somewhat flatter top, suggesting accents other than RP can be heard amongst those who belong to the highest social class (see figure 6.3 on p. 141). In other speech communities, it is certainly possible to hear more than just one accent associated with the highest social group. Most well-educated Scots, Irish and Welsh speakers do not use RP, and there is more than one socially prestigious accent in these countries. And in ex-colonies of Britain such as Australia and Canada, other accents have displaced RP from its former position as the most admired accent of English. In fact, RP now tends to be perceived by many people as somewhat affected (or ‘real posh’!). This kind of negative reaction to RP has also been reported in Britain, especially among young Londoners, many of whom use an accent popularly labeled ‘Estuary English’ when it first attracted attention in the 1990s, because it is claimed to have developed along the Thames Estuary. Some have labelled it the ‘new RP’. In fact, like previous London-based linguistic innovations, features of this modern urban dialect are rapidly spreading both regionally and socially and it is a good illustration of the process of levelling. Levelling involves the reduction of dialect and/or accent variation, and the rapid spread of so-called ‘Estuary English’ has certainly reduced regional variation in a large area of the south of England. It shares many linguistic features with Cockney, another variety which delineates both a region and a social group, but speakers of ‘Estuary English’ don’t usually drop their aitches or use [f] at the begin- ning of the word ‘think’. Social dialects The stereotypical ‘dialect’ speaker is an elderly rural person who is all but unintelligible to modern city dwellers. But the term dialect has a wider meaning than this stereotype suggests. Dialects are linguistic varieties which are distinguishable by their vocabulary, grammar and pronunciation; the speech of people from different social, as well as regional, groups may differ in these ways. Just as RP is a social accent, so standard English is a social dialect. It is the dialect used by well-educated English speakers throughout the world. It is the variety used for national news broadcasts and in print, and it is the variety generally taught in English-medium schools. Standard English Example 8 (a) I’ve not washed the dishes yet today. (b) I haven’t washed the dishes yet today. Standard English is more accommodating than RP and allows for some variation within its boundaries. This is represented in figure 6.3 by the flat top of the trapezium or table-topped mountain. The flat top symbolises the broader range of variants (alternative linguistic forms) 140  Chapter 6 Regional and social dialects Figure 6.3 Social and regional dialect variation Source: From Peter Trudgill (1983) Sociolinguistics. Penguin: London. Page 41. Copyright © Peter Trudgill, 1974, 1983, 1995, 2000. Reproduced by permission of Penguin Books Ltd. which qualify as part of the standard dialect of English in any country. It is estimated that up to 15 per cent of British people regularly use standard British English. So in standard English, a limited amount of grammatical variation is acceptable. A speaker of standard English might produce either of the sentences in example 8 above. The dialect we grace with the name standard English is spoken with many different accents. But, as illustrated in the discussion of regional dialects, there are also many standard Englishes. US standard English is distinguishable from South African standard English and Australian standard English, for instance, and all three differ from the British standard dialect. In social terms, linguistic forms which are not part of standard English are by definition non-standard. Because the standard dialect is always the first to be codified, it is difficult to avoid defining other dialects without contrasting them with the standard. And then, because such non-standard forms are associated with the speech of less presti- gious social groups, the label inevitably acquires negative connotations (though, as discussed in chapter 14, some non-standard dialects, associated with particular groups, are positively evaluated by their users and by those who admire them). It is very important to understand, however, that there is nothing linguistically inferior about non-standard forms. They are simply different from the forms which happen to be used by more socially prestigious speak- ers. To avoid the implication that non-standard forms are inadequate deviations from the standard, some sociolinguists use the term vernacular as an alternative to non-standard, and I will follow this practice. Vernacular is a term which is used with a variety of meanings in sociolinguistics, but the meanings have something in common. Just as vernacular languages contrast with standard languages, vernacular dialect features contrast with standard dialect features. Vernacular forms tend to be learned at home and used in informal contexts. So all uses of the term vernacular share this sense of the first variety acquired in the home and used in casual contexts. Vernacular dialects, like vernacular languages, lack public or overt prestige, though they are generally valued by their users, especially as means of expressing solidarity and affective meaning – a point I will return to in the discussion of attitudes to language in chapter 14. So far I have been discussing accents and dialects as if the linguistic features which identify them were stable, fixed and absolute. But, as with the notion of distinguishable languages, this is just a convenient fiction. The way people speak is characterised by patterned variation. The patterns are fascinating and indicate the social factors which are significant in a society. 141 An introduction to sociolinguistics To illustrate this point, the rest of this chapter will discuss the relationship between speech and social status or class. Caste dialects People can be grouped together on the basis of similar social and economic factors. Their language generally reflects these groupings – they use different social dialects. It is easiest to see the evidence for social dialects in places such as Indonesia and India where social divisions are very clear-cut. In these countries, there are caste systems determined by birth, and strict social rules govern the kind of behaviour appropriate to each group. The rules cover such matters as the kind of job people can have, who they can marry, how they should dress, what they should eat, and how they should behave in a range of social situations. Not surprisingly, these social distinctions have corresponding speech differences. A person’s dialect is an indication of their social background. There are quite clear differences in Indian languages, for example, between the speech of the Brahmins and non-Brahmin castes. The Brahmin word for ‘milk’ in the Kannada lan- guage, for instance, is haalu, while non-Brahmin dialects say aalu. The Tamil Brahmin word for ‘sleep’ is tuungu, while non-Brahmin dialects use the word orangu. And in Tulu, the Brahmin dialect makes gender, number and person distinctions in negative tenses of the verb which are not made in non-Brahmin dialects. In Javanese, too, linguistic differences reflect very clear-cut social or caste divisions. Example 9 An Indonesian student at a British university was trying to explain to her English friend the complications of social dialects in Java and the ways in which Javanese speakers signal their social background. ‘It is much harder than in English,’ she said. ‘It is not just a matter of saying sofa instead of couch, or house rather than ’ouse. Every time you talk to a different person you have to choose exactly the right words and the right pronunciations. Almost every word is different and they fit together in patterns or levels, depending on who you are talking to. Because I am well-educated and come from a rich family, I am expected to use five different levels of language.’ Javanese social status is indicated not just in choice of linguistic forms but also in the par- ticular combinations of forms which each social group customarily uses, i.e. the varieties or stylistic levels that together make up the group’s distinctive dialect. In English, stylistic variation involves choices such as ta mate vs thank you so much. In Javanese, things are very complicated. There are six distinguishable stylistic levels. Table 6.1 provides a couple of words from each level to show the overlap and intermeshing of forms involved. (This example is discussed further in chapter 10 where the reasons for the numbering system are made clear.) There are three distinct Javanese social groups and three associated dialects (see table 6.1). 1. The dialect of the lowest status group, the peasants and uneducated townspeople, consists of three stylistic levels: 1, 1a and 2. 2. The dialect of urbanised people with some education consists of five stylistic levels: 1, 1a, 2, 3 and 3a. 142  Chapter 6 Regional and social dialects Table 6.1 Two Javanese words at different stylistic levels ‘You’ ‘Now’ Stylistic level padjenengan samenika 3a sampéjan samenika 3 sampéjan saniki 2 sampéjan saiki 1a pandjenengan saiki 1a kowé saiki 1 Source: Adapted from The Religion of Java. The Free Press of Glencoe (Geertz, Clifford, 1960). With the permission of The Free Press, a division of Simon & Schuster Inc. Copyright © 1960 renewed © 1988 by Clifford Geertz, all rights reserved. 3. The dialect of the highly educated highest status group also consists of five levels, but they are different from those of the second social group: 1, 1a, 1b, 3 and 3a. In Javanese, then, a particular social dialect can be defined as a particular combination of styles or levels each of which has its distinctive patterns of vocabulary, grammar and pronunciation, though there are many forms which are shared by different stylistic levels. Social class dialects Vocabulary The term social class is used here as a shorthand term for differences between people which are associated with differences in social prestige, wealth and education. In most societies, bank managers do not talk like office cleaners, and lawyers do not speak in the same way as the burglars they defend. Class divisions are based on such status differences. Status refers to the deference or respect people give someone – or don’t give them, as the case may be – and status generally derives in Western society from the material resources a person can command, though there are other sources too. Family background may be a source of status independ- ently of wealth (in Britain, the youngest child of an earl may be poor but respected!). So class is used here as a convenient label for groups of people who share similarities in economic and social status. Social dialect research in many different countries has revealed a consistent relationship between social class and language patterns. People from different social classes speak differently. The most obvious differences – in vocabulary – are in many ways the least illuminating from a sociolinguistic point of view, though they clearly capture the public imagination. In the 1950s in England, many pairs of words were identified which, it was claimed, distinguished the speech of upper-class English people (‘U speakers’) from the rest (‘non-U speakers’). U speakers used sitting room rather than lounge (non-U), and referred to the lavatory rather than the (non-U) toilet. The following excerpt from a Nancy Mitford novel provides an illustration. Example 10 Uncle Matthew: ‘I hope poor Fanny’s school (the word school pronounced in tones of withering scorn) is doing her all the good you think it is. Certainly she picks up some dreadful expressions there.’ ▲ 143 An introduction to sociolinguistics Aunty Emily, calmly, but on the defensive: ‘Very likely she does. She also picks up a good deal of education.’ Uncle Matthew: ‘Education! I was always led to believe that no educated person ever spoke of notepaper, and yet I hear poor Fanny asking Sadie for notepaper. What is this education? Fanny talks about mirrors and mantlepieces, handbags and perfume, she takes sugar in her coffee, has a tassel on her umbrella, and I have no doubt that, if she is ever fortunate enough to catch a husband, she will call his father and mother Father and Mother. Will the wonderful education she is getting make up to the unhappy brute for all these endless pinpricks? Fancy hearing one’s wife talking about notepaper – the irritation!’ Aunty Emily: ‘A lot of men would find it more irritating to have a wife who had never heard of George III. (All the same, Fanny darling, it is called writing paper you know – don’t let’s hear any more about note, please.)’ If these vocabulary differences exist at all, they are rather like those which distinguish Brahmin and non-Brahmin castes – they distinguish social groups on a categorical basis. You either use the U term or you don’t! It is not quite so simple, of course, since inevitably the U terms spread beyond the boundaries of the U group and so new terms are introduced. By the 1970s, only non-U speakers, it was claimed, used ‘handbag’ for U ‘bag’, a non-U ‘settee’ was a U ‘sofa’, and your non-U ‘relatives’ were U ‘relations’. There is no empirical research to back up these claims, but even if they exist, vocabulary clues are superficial and conceal the complexity and relative fluidity of social class membership in places like Britain, and even more so in the USA. The barriers between groups are not insurmountable as in caste-based societies. People can move up or down the social ladder, and this potential mobility is mirrored more accurately in other aspects of their speech – such as pronunciation. Exercise 6 A preference for different vocabulary by different social groups is relatively easy to identify and always fascinates people. Can you produce a list of words for your speech community that divides people up according to their social background? Words for death, lavatory, really good and under the influence of alcohol often vary from one social group to another. Pronunciation Example 11 Kim: Only uneducated people drop their ‘h’s. Stephen: Let’s hear you say ‘Have you heard about Hilda’s new house that her husband left her? It cost her a heck of a lot to fix up.’ If you don’t drop a single ‘h’ in that sentence you’ll sound like one of Monty Python’s upper-class twits! 144  Chapter 6 Regional and social dialects In describing differences between Canadian and Parisian French, I mentioned that the differ- ences are often not absolute, but rather matters of frequencies. Exactly the same is true for the speech of different social groups. Groups are often distinguished by the frequency with which they use particular features, rather than by their use of completely different forms. This important point can be illustrated with a simple example. Usha Pragji, a New Zealand student, taped a radio broadcast of two elderly people’s accounts of their childhood in Edwardian Wellington, the capital city of New Zealand. The two speakers contrasted on a range of social variables. Example 12 Marjorie Lee lived in what she described as ‘a new large hideous Edwardian mansion’. Her father was a lawyer. Her mother had been one of the first women students at Canterbury College, part of the University of New Zealand. Marjorie went to Miss Barber’s Private School until her early teens when she was provided with a governess. George Davies lived with four members of his family in a small house. His mother was a solo parent whose only steady income came from ‘keeping grandmother’. As George said, ‘we weren’t hard-up, we were absolutely poverty-stricken.’ Illustrating this, he recounted that they sometimes took palings off the fence for firewood, and ‘you could pretty well tell the state of the family finances by whether Grandma’s inlaid brooch was in the pawn shop or not.’ The two recordings were analysed to see whether there were differences between the speakers in terms of the numbers of [h]s they ‘dropped’ in words like house. Usha found that Marjorie Lee did not omit a single [h] while George Davis dropped 83 per cent of the [h]s which occurred in his interview. The speakers’ different social backgrounds were clearly signalled by the different proportions of this feature of their speech. This speech variable is widely called [h]-dropping – a label which you should note represents the viewpoint of speakers of the standard. It has been analysed in many social dialect studies of English. Figure 6.4 shows the average [h]-dropping scores for five different social groups in two different places in England: West Yorkshire and Norwich. Different studies use different labels for different classes so wherever I compare data from different studies I have simply used numbers. In this figure (and throughout the book), group 1 refers to the highest social group (often called the upper middle class or UMC) and group 5 to the lowest (usually called the lower or working class). In both areas the highest social group drops the least number of [h]s and the lowest group omits the most. The regularity of the patterns from one group to the next seems quite remark- able when you first see a diagram of this kind, but it is quite typical. The average scores for socially differentiated groups generally follow exactly this kind of pattern for any linguistic variable where there are clearly identifiable standard and vernacular variants. There are regional differences in that the West Yorkshire scores are systematically higher than the Norwich scores, but the overall pattern remains the same. We also need to remember that these are averages and within each social group there is always a great deal of individual variation. In the West Yorkshire study, for example, one person who belonged socially in the middle group (3) dropped every [h]. From a linguistic point of view, taking account only of 145 An introduction to sociolinguistics Figure 6.4 [h]-dropping in Norwich and West Yorkshire social groups Source: This diagram was constructed from data in Trudgill (1974) and Petyt (1985). [h]-dropping, they sounded as if they came from a lower social group. Averaging may conceal considerable variation within a group. The way different pronunciations fall into a pattern indicating the social class of their speakers was first demonstrated by William Labov in a study of New York City speech which is now regarded as a classic in sociolinguistics. He designed a sociolinguistic interview to elicit a range of speech styles from 120 people from different social backgrounds, and then he analysed their pronunciations of a number of different consonants and vowels. He found regular patterns relating the social class of the speakers to the percentage of standard as opposed to vernacular pronunciations they produced. Some of the linguistic features he studied have been found to pattern socially in English-speaking communities all over the world. The pronunciation -ing vs -in’ ([ih] vs [in]) at the end of words like sleeping and swimming, for instance, distinguishes social groups in every English-speaking community in which it has been investigated. The figures in table 6.2 demonstrate that, as with [h]-dropping, there are regional variations between communities, but the regularity of the sociolinguistic pattern in all four communities is quite clear.1 The Brisbane data was collected from adolescents, but the data from the other com- munities is representative of the communities as a whole. In each community, people from lower social groups use more of the vernacular [in] variant than those from higher groups. Table 6.2 Percentage of vernacular [in] pronunciation for four social groups in speech communities in Britain, the USA and Australia Social group 1 2 3 4 Norwich 31 42 91 100 West Yorkshire 5 34 61 83 New York 7 32 45 75 Brisbane 17 31 49 63 146  Chapter 6 Regional and social dialects Example 13 In New York City in 1964, a man was observed in three different department stores asking one store worker after another: ‘where are the women’s shoes?’ The man appeared not only to have a short memory, since he repeated his question to a shop assistant in each aisle on several different floors, he also appeared to be slightly deaf since he asked each person to repeat their answer to him. After receiving the answer he would scurry away and scribble something in his notebook. Oddest of all, when he finally made it to the fourth floor where the women’s shoes were, he showed absolutely no interest in them whatsoever but wandered around the floor asking, ‘Excuse me, what floor is this?’ When questioned by a puzzled store detective, he said he was a sociolinguist! One linguistic form which has proved particularly interesting to sociolinguists studying English-speaking speech communities is the variable pronunciation of [r] in words like car and card, for and form. For our purposes, there are two possible variants of [r]. Either it is present and pronounced [r], or it is absent. If you listen to a range of dialects you will find that sometimes people pronounce [r] following a vowel, and sometimes they don’t. In some regions, pronouncing [r] is part of the standard prestige dialect – in the Boston and New York areas of the eastern USA, for example, in Ireland and in Scotland (though recent research suggests that this is changing in the speech of young working-class people in Glasgow and Edinburgh). In other areas, standard dialect speakers do not pronounce [r] after vowels (or ‘post-vocalically’ as linguists describe it) in words like car and card.2 In areas where [r] pronunciation is prestigious, sociolinguists have found patterns like those described above for [h]-dropping and -in’ vs -ing ([in] vs [ih]) pronunciation. The higher a person’s social group, the more [r] they pronounce. In New York City, Labov conducted an interesting experiment demonstrating in a neat and economical way that pronunciation of post-vocalic [r] varied in the city according to social group. As example 13 describes, he asked a number of people in different department stores where to find an item which he knew was sold on the fourth floor. Then, pretending he hadn’t heard the answer, he said, ‘Excuse me?’ People repeated their answers and he obtained a second and more careful pronunciation. So each person had the chance to pronounce [r] four times: twice in fourth and twice in floor. This ingenious rapid and anonymous survey technique provided some interesting patterns. The results showed clear social stratification of [r] pronunciation. Overall, the ‘posher’ the store, the more people used post-vocalic [r]. And even within stores a pattern was evident. In one store, for instance, nearly half the socially superior supervisors used post-vocalic [r] consistently, while only 18 per cent of the less statusful salespeople did, and the stock boys rarely used it at all. Post-vocalic [r] illustrates very clearly the arbitrariness of the particular forms which are considered standard and prestigious. There is nothing inherently bad or good about the pronunciation of any sound, as the different status of [r]-pronunciation in different cities illustrates. In New York City, pronouncing [r] is generally considered prestigious. In Reading in England it is not. This is apparent from the patterns for different social groups in the two cities illustrated in figure 6.5. In one city the higher your social class the more you pronounce post-vocalic [r]. In the other, the higher your social class the fewer you pronounce. 147 An introduction to sociolinguistics Figure 6.5 Post-vocalic [r] in Reading and New York social group Source: This diagram was constructed from data in Romaine (1984: 86). Example 14 Cholmondley: Which ’otel are you staying at old chap? The same point is illustrated by [h]-dropping. Kim’s comment in example 11 expresses a widely held viewpoint – only uneducated people drop their [h]s. Some claim that [h]-dropping is evidence of laziness or slovenly speech. Yet well into the twentieth century, the top social classes in England dropped the [h] at the beginning of words like hotel and herb. Interestingly, initial [h] has recently reappeared in the speech of young Londoners who belong to ethnic minority groups. Clearly, the particular linguistic forms which people regard as prestigious or stigmatised are in general totally arbitrary. In New York City, older white residents of the Lower East Side use non-rhotic pronunciations to assert their status as authentic New Yorkers. The arbitrariness of the particular sounds invested with social significance is probably most obvious in the pronunciation of vowels where the precise values of the standard forms are entirely determined by the speech of the most prestigious social group. Example 15 Sir – What is happening to the humble letter ‘i’ in New Zealandese? In many mouths HIM becomes HUM, JIM is JUM and TILL is TULL. I overheard a young girl telling her friend on the phone that she had been to a doctor and had to take six different PULLS a day. After four repetitions, she had to spell it to be understood. 148  Chapter 6 Regional and social dialects Source: © 1986 United Feature Syndicate, Inc. Reproduced by permission. In this letter, the writer is complaining about the way New Zealanders pronounce their vowels. The examples illustrate the difficulty of precisely describing differences in the pro- nunciation of vowels without the aid of a phonetic script. Measuring slight differences in the way speakers pronounce the ‘same’ vowels is also a challenging task. In his pioneering study of New York speech, Labov measured people’s pronunciation of five vowels as well as a number of consonants. Measuring the presence or absence of [h] or [r], or the difference between [in] and [ih], is difficult enough when you are listening to tapes of interviews. Measuring small but significant differences in vowel pronunciations can seem a nightmare. Labov developed a method which involved giving a score to different pronunciations according to how close they were to the prestige pronunciation or standard in the community. The scoring system is most easily understood by giving an example. In New Zealand, a survey of 141 people living in the South Island distinguished three different social groups on the basis of the way speakers pronounced the diphthongs (gliding vowels) in words such as boat, bite and bout. Four points on a scale were used to measure the different pronunciations. A score of 4 was allocated to pronunciations closest to RP, and a score of 1 to the ‘broadest’ New Zealand pronunciations, with two points in between. A pattern emerged with the highest social group scoring 60 or more (out of a possible 100) for these diphthongs, the middle group scoring between 50 and 55, while the lowest social group scored less than 43 (where 25 was the minimum possible score). Many New Zealanders consider RP an inappropriate standard accent for New Zealand but, at least in the late twentieth century when this study was undertaken, it was still an influential prestige norm. The systematically patterned scores for diphthong pronunciation in different social groups clearly revealed the social basis of New Zealand patterns of pronunciation. The higher a person’s social class, the closer their pronunciation was to RP. The scoring system allowed comparison of the way different social groups pronounced these vowels. Exercise 7 You will find that you can collect examples yourself to illustrate the patterns described, though it takes a bit of practice to accurately identify the sounds you are listening for. Audio-record the best-educated person you know, and the person with the least education. Ask them to describe the first school they went to. Then count the number of [iy]s they use compared to the number of [in]s at the end of words like swimming and running. Check whether any differences you find between your speakers in the numbers of [in] vs [iy] are consistent with the findings reported in this chapter. 149 An introduction to sociolinguistics Example 16 Jean Charmier is a young Parisian who works as a labourer on a construction site. His speech is quite different in many ways from the speech of the news announcers on national television. One difference involves his pronunciation of the as in words like casser (‘break’) and pas (‘not’). He says something which sounds more like cosser for casser and instead of je ne sais pas for ‘I don’t know’, he omits the ne (as most people now do in colloquial French) and pronounces the phrase as something which sounds like shpo. Although the sociolinguistic patterns described in the preceding section have been most extensively researched in English-speaking communities, they have been found in other languages too. In fact we would expect to find such patterns in all communities which can be divided into different social groups. In Paris, the pronunciation of the first vowel in words like casser and pas varies from one social group to another. In Montreal, the frequency with which [l] is deleted distinguishes the French of two social groups, as illustrated in table 6.3. Table 6.3 Percentage [l]-deletion in two social classes in Montreal French Professional Working Class il (impersonal) 89.8 99.6 e.g. il pleut ‘it is raining’ il (personal) 71.6 100.0 e.g. il part ‘he is leaving’ elle 29.8 82.0 Source: Reproduced from Sankoff and Cedergren 1971: 81. This table also introduces another interesting influence on linguistic variation. The pro- nunciation of a linguistic form often alters in different linguistic contexts. Table 6.3 shows that not only does [l]-deletion differ between the social classes, it also differs according to the grammatical status of the word in which it occurs. [l] has almost disappeared in Montreal French in impersonal il. The surrounding sounds also affect the likelihood of [l]-deletion. It is much more likely to disappear before a consonant than before a vowel. This is the point being made by Stephen in example 11. Most people ‘drop their hs’ in an unstressed syllable. So linguistic as well as social factors are relevant in accounting for patterns of pronunciation. But within each linguistic context the social differences are still quite clear. Similar patterns can be found in any speech community where there is social stratification. In Tehrani Persian, as well as in the Swahili used in Mombasa, the same relationship is found between speech and social class. The higher social groups use more of the standard forms, while the lowest groups use the fewest standard forms. 150  Chapter 6 Regional and social dialects Grammatical patterns Example 17 Whina is 8 years old and she is telling a visitor the story of a film she has seen. ‘And then these little flies went to go and they made a house by theirself, and this big fly was playing his guitar. He play and play. Then the little flies was making the house, and then the flies um sew um these leaves up all together.’ This is an extract from a recorded interview. Whina was one of eighty New Zealand children who were recorded telling the story of a film they had seen. As the extract shows, this was a good way of providing a natural context for the children to use a large number of past tense forms of verbs. It was then possible to compare the proportion of standard verb forms in the speech of children from different social groups. On average, it was found that children from lower-class families used more vernacular verb forms than children from middle-class families. This pattern has been noted for a variety of grammatical variables. Here are some examples of standard and vernacular grammatical forms which have been identified in several English- speaking communities. Form Example Past tense verb forms 1. I finished that book yesterday. 2. I finish that book yesterday. Present tense verb forms 3. Rose walks to school every day. 4. Rose walk to school every day. Negative forms 5. Nobody wants any chips. 6. Nobody don’t want no chips. Ain’t 7. Jim isn’t stupid. 8. Jim ain’t stupid. As with pronunciation, there is a clear pattern to the relationship between the grammatical speech forms and the social groups who use them. Figure 6.6 illustrates this. The higher social groups use more of the standard grammatical form and fewer instances of the vernacular or non-standard form. With the grammatical pattern illustrated in figure 6.5, the third person singular form of the present tense regular verb (e.g. standard she walks vs vernacular she walk), there is a sharp distinction between the middle-class groups and the lower-class groups. Sociolinguists describe this pattern as sharp stratification. People are often more aware of social stigma in relation to vernacular grammatical forms, and this is reflected in the lower incidence of vernacular forms among middle-class speakers in particular. Note that this pattern is found both in a variety of US English spoken in Detroit, and in a variety of British English spoken in Norwich. Sentence (6) in the list illustrates a pattern of negation which is sometimes called ‘negative concord’ or ‘multiple negation’. Where standard English allows only one negative in each clause, most vernacular dialects can have two or more. In some dialects, every possible form which can be negated is negated. An adolescent gang member in New York produced the following: 151 An introduction to sociolinguistics Figure 6.6 Vernacular present tense verb forms (3rd person singular: she walk) in Norwich and Detroit Source: This diagram was constructed from data in Trudgill (1983) and Wolfram and Fasold (1974). 9. It ain’t no cat can’t get in no coop. Translated into standard English, the meaning of this utterance in context was 10. There isn’t any cat that can get into any (pigeon) coop or, more simply, no cat can get into any coop. Sentence (11) comes from an adolescent in Detroit: 11. We ain’t had no trouble about none of us pulling out no knife. Multiple negation is a grammatical construction which has been found in all English-speaking communities where a social dialect study has been done. In every community studied, it is much more frequent in lower-class speech than in middle-class speech. In fact, there is usually a dramatic contrast (sharp stratification again) between the groups in the amount of multiple negation used. It is rare in middle-class speech. Multiple negation is a very ‘salient’ vernacular form. People notice it when it is used even once, unlike say the use of a glottal stop for the standard pronunciation [t] at the end of a word, where the percentage of glottal stops generally needs to be quite high before people register them. The dramatic split evident between middle-class and lower-class usage of multiple negation indicates this salience. Middle-class speakers tend to avoid it, while lower-class speakers use it more comfortably. In reporting patterns relating linguistic features to social status, I have inevitably simplified a great deal. Many factors interact in determining the proportion of vernacular or standard forms a person uses. Some of these are social factors such as the age or gender of the speaker, and they will be examined in the next couple of chapters. Another factor, however, which was mentioned briefly above, is the linguistic environment in which a word occurs. Table 6.3, for example, provided information on the effect of the linguistic environment in which [l]- deletion in French occurred. In some varieties of English, it has been found that people omit 152  Chapter 6 Regional and social dialects the auxiliary verb have more often before got (e.g. you got to go) than before other verbs (e.g. you’ve done that). The lexical context appears to be relevant. Before analysing the effects of social variation, then, social dialectologists must take into account the effects of the linguistic environment in which the linguistic feature occurs. Exercise 8 (a) Make three points about the distribution of non-standard or vernacular forms in British urban dialects which are supported by figure 6.7. (b) This diagram shows values for 3 towns in England. Similar diagrams can be drawn for social classes. If the labels ‘Hull’, ‘Milton Keynes’ and ‘Reading’ referred to social classes, which would be the lowest social class? How do you know? Figure 6.7 Vernacular forms in three English towns Source: Cheshire, J. (1999: 129–48). Note: vernacular ‘was’ = use of was where standard English uses were: e.g. you was late again. Negative concord = multiple negation. Milton Keynes, Reading and Hull are British towns. Milton Keynes is a relatively new town 80 kilometres north of London. Reading is 60 kilometres west of London and Hull is the furthest north and over 200 kilometres from London. Source: Cheshire, Kerswill and Williams (2005). Reproduced with permission. Answers at end of chapter A note about methodology Example 18 My friend Terry taught sociolinguistics at a New Zealand university. He described to his students how I and my colleagues Allan Bell and Mary Boyce designed our sample for the Wellington social dialect survey by establishing a quota of 5 cells for each of our target groups: young Pakeha women, middle-aged Maori men and so on. After the exams, he rang to tell me that some of his students had clearly misunderstood our method of collecting data. In their exam answers, they described how we had put people in cells before interviewing them! 153 An introduction to sociolinguistics Map 6.1 Map of England showing distance between London and other cities The term ‘cell’ is used to describe the box in a diagram which indicates the target numbers for each group in a sample. Our goal was a minimum of 5 people in each box or cell in the diagram to represent each group, including women and men, three age groups, two ethnic groups (Maori and Pakeha) and two social groups. We managed to reach our target, but not by capturing people and putting them in cells till they agreed to be interviewed! Table 6.4 illustrates a sample of 20 people divided by age and gender with a quota of five people per cell. Table 6.4 A sample of 20 by age and gender with a quota of 5 people per cell Age Young Middle-aged Gender Male 5 young males 5 middle-aged males Female 5 young females 5 middle-aged females Collecting good-quality social dialect data requires considerable skill. Some of the challenges and pitfalls which face the social dialectologist will emerge during the discussion in the next few chapters. In this chapter, we have mentioned the use of rapid and anonymous surveys as one useful technique for collecting a lot of data very quickly. Labov used this method when he asked ‘where are the women’s shoes?’ in the New York Department store. Another example is the use of a street survey to ask people to pronounce certain words (ideally presenting them in written or visual form to prevent possible bias from the interviewer’s pronunciation). This method allows the researcher to access features of the speech of a large number of people in a 154  Chapter 6 Regional and social dialects relatively short time, and can provide interesting clues about what is worth studying in more detail. But it has obvious limitations. Though patterns of gender and age and sometimes ethnic variation can be detected using this approach, since we can usually guess these features when we meet someone, it is harder to be sure about the social background of speakers when no social information is collected from them. And, of course, the data is limited in quantity and style. A sociolinguistic interview which attempts to elicit a range of styles as well as collect background information from the interviewee is thus much more useful. This is by far the most widespread method of collecting social dialect data. But of course it is much more time- consuming and thus expensive. Conclusion The way you speak is usually a good indicator of your social background. And there are many speech features which can be used as clues. Sociolinguists have found that almost any linguistic feature in a community which shows variation will differ in frequency from one social group to another in a patterned and predictable way. Some features are stable and their patterns of use seem to have correlated with membership of particular social groups in a predictable way for many years. Variation in pronunciation of the suffix -ing and [h]-dropping are examples of features which are usually stable. Grammatical features, such as multiple negation and tense markers, are often stable too. This means they are good ones to include in any study of an English-speaking community. They are reliable indicators of sociolinguistic patterning in a community. Social dialect surveys have demonstrated that stable variables tend to divide English-speaking communities sharply between the middle class and lower or working classes. So patterns of [h]-dropping and [ih] vs [in] pronunciation clearly divide the middle-class groups from the lower-class groups in Norwich. Grammatical variables do the same, as figure 6.5 on p. 148 illustrated. There is a sharp rise in the number of vernacular forms between the middle-class groups and the lower-class groups both in Detroit and in Norwich, and the same pattern has been observed in many other communities. As mentioned above, this pattern has been labelled sharp stratification. Not all variation is stable over time, however. In fact, variation is often used as an indicator of language change in progress. New linguistic forms don’t sweep through a community over- night. They spread gradually from person to person and from group to group and they often stratify the population very delicately or finely. Unstable variation is thus associated with fine stratification and is a clue to linguistic changes in progress in the community, as we will see in chapter 9. I have focused largely on pronunciation and grammar in this chapter, but social dialectologists are increasingly paying attention to pragmatic features too. The way people use tag questions (isn’t she, didn’t they), for example, or pragmatic particles such as you know, may also index their social group. In exploring the relationship between language and society, this chapter has been con- cerned almost exclusively with the dimension of social status or class. The evidence discussed indicates that the social class someone belongs to is generally signalled by their speech patterns. Many people, however, are not very conscious of belonging to a particular social class. They are much more aware of other factors about the people they meet regularly than their social class membership. A person’s gender and age are probably the first things we notice about them. The next chapter explores ways in which women and men speak differently, and describes speech differences associated with age. 155 An introduction to sociolinguistics Answers to exercises in chapter 6 Answer to exercise 2 Using the system for representing sounds in Appendix 1, your answer for batter might include: [bata] the pronunciation considered standard in Britain [batar] a pronunciation considered standard in North America [baca] a Cockney pronunciation [bacar] and [badar] pronunciations heard in the West Country of England There are many other possible pronunciations, including one common in Liverpool which sounds a little like [batsa] with affrication after the [t], and another which occurs in New Zealand English which sounds like [beta] to British ears. Answers to exercise 3 (a) ‘snowflake’ (Somerset) (b) ‘time for a snack’ (Norfolk and elsewhere too) (c) ‘infuse or brew the tea’ (Cumberland, Durham, Northumberland, Yorkshire, Scotland) (d) ‘the place was all untidy, disordered’ (Lincolnshire) (e) ‘icicles’ (Somerset) (f ) ‘a great rough awkward girl’ (Essex, East Anglia and elsewhere) (g) ‘I’m really fed up, weary’ (Yorkshire) (h) ‘child’ (Scottish) (i) ‘newt’ (south-east) (j) a piggy is a ‘hot water bottle’ (Scotland, Northumberland) Answers to exercise 5 (a) Accents are distinguished from each other by pronunciation alone. Different dialects are generally distinguishable by their pronunciation, vocabulary and grammar. (b) Regional dialects involve features of pronunciation, vocabulary and grammar which differ according to the geographical area the speakers come from. Social dialects are dis- tinguished by features of pronunciation, vocabulary and grammar according to the social group of the speakers. Social group is usually evaluated on the basis of a range of features, such as education, occupation, residential area and income level. So people who come from different social groups speak different social dialects if they use different words, pro- nunciations, and grammatical features. Examples of these are discussed in the next section. Answers to exercise 8 (a) Example of points you might have made: the pattern of vernacular usage is consistent for both variables in all 3 places negative concord is slightly more frequent than non-standard was in 2 of the 3 places vernacular was is particularly distinctive of Hull speech, where it reaches almost 100 per cent or categorical status the further north you go the higher the percentage of vernacular forms this table suggests that regional variation intersects with class variation since, as the section preceding this exercise indicates, these vernacular forms typically stratify urban populations socially. 156  Chapter 6 Regional and social dialects (b) Hull would be the equivalent of the lowest social class because the percentages of vernacular forms are highest in Hull and lower social groups tend to use more vernacular forms. Concepts introduced Accent Regional dialect Isogloss Dialect chain Social dialect Vernacular Sociolinguistic patterns Methodology Sharp and fine stratification References The following sources provided material for this chapter: Bayard (1987), Bayard et al. (2001) for New Zealand data Becker (2009) on New York City and non-rhoticity Bright (1966) on Indian languages Bright and Ramanujan (1964) on Indian languages Chambers and Trudgill (2004) Cheshire, Kerswill and Williams (2005) is the source of figure 6.7 Eisikovits (1989a) on have deletion Feagin (1979) for data on Anniston and West Virginia Finegan and Besnier (1989: 383) for definition of a language Gordon and Deverson (1985, 1989) Guy (1988: 37) on ‘social class’ Kerswill (2001) on dialect levelling and Estuary English Labov (1966, 1972a, 1972b), especially data on New York City Lee (1989) on Brisbane adolescents’ speech McCallum (1978) on New Zealand past tense usage Mitford and Ross (1980) and Wales (1994) on ‘U’ speech Orton, Sanderson and Widdowson (1998) on British English dialects Petyt (1985) for West Yorkshire dialect data Pragji (1980) on [h] dropping Romaine (1978, 1984) on Edinburgh speech Sankoff and Cedergren (1971) on Montreal French Shuy, Wolfram and Riley (1967) for Detroit data Stuart-Smith, Timmins and Tweedie (2007) on Glasgow speech Trudgill (1974, 1983), especially data on Norwich Wolfram and Fasold (1974) for Detroit data Quotations Example 10 is from Mitford (1949: 29) Example 15 is from Gordon and Deverson (1989: 35) 157 An introduction to sociolinguistics Useful additional reading Chambers and Trudgill (2004) Coupland and Jaworski (2009) Part 1: Language Variation Hughes and Trudgill (2005) Mesthrie et al. (2009) Chs 3 and 4 Milroy and Gordon (2003) Tagliamonte (2006) Trudgill (2000), Ch. 2 Wardhaugh (2010), Chs 2 and 7 Wolfram and Schilling-Estes (2006) Notes 1. Here and elsewhere I have combined data from groups 3 (upper working class) and 4 (middle working class) in Norwich and West Yorkshire for ease of comparison with communities which were analysed into only four social groups. 2. ‘Post-vocalic’ is the term widely used for the [r] in words like car and card which is not pronounced in many varieties of English, and I use this term throughout the book. Note, however, that strictly speaking one should refer to non-pre-vocalic (r) in such contexts, since [r] is always pronounced between vowels in words such as wary and carry. 158  7 Gender and age D o women and men speak differently? Do children speak differently from adults? The answer to both these questions is almost certainly ‘yes’ for all speech communities, and the reasons in both cases are mainly social and cultural. The linguistic forms used by women and men contrast – to different degrees – in all speech communities. There are other ways too in which the linguistic behaviour of women and men differs. It is claimed women are more linguistically polite than men, for instance, and that women and men emphasise different speech functions. These claims will be explored in later chapters. In the first section of this chapter, the focus will be on evidence that women and men from the same speech community may use different linguistic forms. First a brief comment on the meaning of the terms sex and gender in sociolinguistics. I have used the term gender rather than sex because sex has come to refer to categories distinguished by biological characteristics, while gender is more appropriate for distinguishing people on the basis of their socio-cultural behaviour, including speech. The discussion of gender in this chapter focuses largely on contrasts between empirically observed features of women’s and men’s speech. The concept of gender allows, however, for describing masculine and feminine behaviours in terms of scales or continua rather than absolute categories. So we can also think of the features associated with women and men’s speech as linguistic resources for constructing ourselves as relatively feminine or relatively masculine. This is something which is discussed further in chapter 12. Gender-exclusive speech differences: highly structured communities Example 1 Tayana is a young Amazonian Indian woman from the north-west Amazon Basin. She lives with her husband and children and a number of other families in a longhouse beside the river. The language of her longhouse is Tuyuka, which is the language of all the men in this tribe, and the language she uses to talk to her children. She comes from a different tribe and her first language is Desano. She uses Desano to her husband, and he replies in Tuyuka. 159 An introduction to sociolinguistics Map 7.1 Colombia and Brazil Women and men do not speak in exactly the same way as each other in any community. The Amazon Indians provide an extreme example. As described in chapter 4, in any longhouse the language used by a child’s mother is different from her father’s language, because men must marry outside their own tribe, and each tribe is distinguished by a different language. In this community, women and men speak different languages. Less dramatically, there are communities where the language is shared by women and men, but particular linguistic features occur only in the women’s speech or only in the men’s speech. These features are usually small differences in pronunciation or word-shape (morphology). In Montana, for instance, there are pronunciation differences in the Gros Ventre American Indian tribe. Where the women say [kjajtsa] for ‘bread’ the men say [dfajtsa]. In this community, if a person uses the ‘wrong’ form for their gender, the older members of the community may consider them bisexual. In Bengali, a language of India, the women use an initial [l] where the men use an initial [n] in some words. Word-shapes in other languages contrast because women and men use different affixes. In Yana, a (now extinct) North American Indian language, and Chiquitano, a South American Indian language, some of the words used between men are longer than the equivalent words used by women and to women, because the men’s forms sometimes add a suffix, as illustrated in example 2. 160  Chapter 7 Gender and age Example 2 Yana Women’s form Men’s form ba ba-na ‘deer’ yaa yaa-na ‘person’ t’et t’en’-na ‘grizzly bear’ cau cau-na ‘fire’ nisaaklu nisaaklu-ci ‘he might go away’ Just the reverse is true for Yanyuwa, an endangered Australian aboriginal language, where it is often the women’s forms which are longer because men and women use different forms of the class-marking prefixes on noun classes, verbs and pronouns. In traditional and con- servative styles of Japanese, forms of nouns considered appropriate for women are frequently prefixed by o-, a marker of polite or formal style. In some languages, there are also differences between the vocabulary items used by women and men, though these are never very extensive. Traditional standard Japanese provides some clear examples. Example 3 Japanese Women’s form Men’s form otoosan oyaji ‘father’ taberu kuu ‘eat’ onaka hara ‘stomach’ In modern standard Japanese, these distinctions are more a matter of degrees of formality or politeness than gender; so the ‘men’s’ forms are largely restricted to casual contexts and are considered rather vulgar, while the ‘women’s’ forms are used by everyone in public contexts. Increasingly, too, as gender roles change, with more women in the workforce and more men prepared to assist in child-rearing, young Japanese women are challenging restrictive social norms, and using the ‘men’s’ forms. While initially women who used these forms were regarded as rather ‘macho’, the social meaning of these forms is changing. They are no longer so much signals of masculinity as of informality and modernity. Some languages signal the gender of the speaker in the pronoun system. In Japanese, for instance, there are a number of words for ‘I’ varying primarily in formality (a point explored further in chapter 10), but women are traditionally restricted to the more formal variants. So ore is used only by men in casual contexts and boku, the next most casual form, is used mainly by men in semi-formal contexts, while women are conventionally expected to use only the semi-formal variant, atashi, the formal watashi and the most formal watakushi (forms also used by men in formal contexts). However, again modern young Japanese women are increasingly challenging such restrictions. 161 An introduction to sociolinguistics Map 7.2 Japan Exercise 1 Do English pronouns encode the gender of the speaker? Answer at end of chapter Gender differences in language are often just one aspect of more pervasive linguistic differences in the society reflecting social status or power differences. If a community is very hierarchical, for instance, and within each level of the hierarchy men are more powerful than women, then linguistic differences between the speech of women and men may be just one dimension of more extensive differences reflecting the social hierarchy as a whole. In Bengali society, for instance, a younger person should not address a superior by first name. Similarly, a wife, being subordinate to her husband, is not permitted to use his name. She addresses him with a term such as suncho ‘do you hear?’ When she refers to him, she uses a circumlocution. One nice example of this practice is provided by the Bengali wife whose husband’s name was tara, 162  Chapter 7 Gender and age which also means ‘star’. Since she could not call him tara, his wife used the term nokkhotro or ‘heavenly body’ to refer to him. This point – the interrelationship of gender with other social factors – is illustrated even more clearly in the next section. The fact that there are clearly identifiable differences between women’s and men’s speech in the communities discussed in this section reflects the clearly demarcated gender roles in these communities. Gender-exclusive speech forms (i.e. some forms are used only by women and others are used only by men) reflect gender-exclusive social roles. The responsibilities of women and men are different in such communities, and everyone knows that, and knows what they are. There are no arguments over who prepares the dinner and who puts the children to bed. Gender-preferential speech features: social dialect research Example 4 Keith was a 7-year-old Canadian from Vancouver whose parents were working for six months in the city of Leeds in Yorkshire, England (see map page 154). He had been enrolled at the local school, and after his first day Keith came home very confused. ‘What’s your teacher’s name?’ asked his father. ‘She says she’s Mrs Hall,’ said Keith, ‘but when the boys call her Mizall she still answers them. And the girls sometimes call her Mrs Hall and sometimes Mizall. It sounds very funny.’ Not surprisingly, in Western urban communities where women’s and men’s social roles overlap, the speech forms they use also overlap. In other words, women and men do not use completely different forms. They use different quantities or frequencies of the same forms. In all the English-speaking cities where speech data has been collected, for instance, women use more -ing [ih] pronunciations and fewer -in’ [in] pronunciations than men in words like swimming and typing. In Montreal, the French used by women and men is distinguished by the frequencies with which they pronounce [l] in phrases such as il y a and il fait. Both women and men delete [l], but men do so more often than women. In Sydney, some women and men pronounce the initial sound in thing as [f], but the men use this pronunciation more than the women. Both the social and the linguistic patterns in these communities are gender- preferential (rather than gender-exclusive). Though both women and men use particular forms, one gender shows a greater preference for them than the other. In all these examples, women tend to use more of the standard forms than men do, while men use more of the vernacular forms than women do. In Australia, interviews with people in Sydney revealed gender-differentiated patterns of [h]-dropping. Exercise 2 What would you predict for [h]-dropping patterns? Is it more likely that women or men drop most [h]s? Answer at end of chapter 163 An introduction to sociolinguistics Gender and social class Example 5 Linda lives in the south of England and her dad is a lawyer. When she was 10 years old, she went to stay for a whole school term with her uncle Tom and auntie Bet in Wigan, a Lancashire town, while her mother was recovering from a car accident. She was made to feel very welcome both in her auntie’s house and at the local school. When she went home, she tried to describe to her teacher what she had noticed about the way her uncle and auntie talked. ‘Uncle Tom is a plumber,’ she told Mrs Button ‘and he talks just like the other men on the building site where he works – a bit broad. He says ’ouse and ’ome and [kup] and [bus]. When she’s at home auntie Bet talks a bit like uncle Tom. She says “Me feet are killin’ me [luv]. I’ve ’ad enough standin’ [up] for today.” But she works in a shop and when she’s talking to customers she talks more like you do Mrs Button. She says house and home and she talks real nice – just like a lady.’ The linguistic features which differ in the speech of women and men in Western commun- ities are usually features which also distinguish the speech of people from different social classes. So how does gender interact with social class? Does the speech of women in one social class resemble that of women from different classes, or does it more closely resemble the speech of the men from their own social class? The answer to this question is quite complicated, and is different for different linguistic features. There are, however, some general patterns which can be identified. In every social class where surveys have been undertaken, men use more vernacular forms than women. Figure 7.1 shows, for instance, that in social dialect interviews in Norwich, men used more of the vernacular [in] form at the end of words like speaking and walking than women. And this pattern was quite consistent across five distinct social groups. (Group 1 represents the highest social group.) Figure 7.1 Vernacular [in] by sex and social group in Norwich Source: This diagram was devised from data in Trudgill (1983). 164  Chapter 7 Gender and age Source: DENNIS THE MENACE ® used by permission of Hank Ketcham Enterprises and © North America Syndicate. Notice, too, that in the lowest and the highest social groups the women’s speech is closer to that of the men in the same group than to that of women in other groups. In these groups, class membership seems to be more important than gender identity. But this is not so true of women in group 2. Their score (of 3 per cent) for vernacular forms is closer to that of women in group 1 than it is to that of men from their own group. This may indicate they identify more strongly with women from the next social group than with men from their own social group. Possible reasons for this are discussed below. 165 An introduction to sociolinguistics Exercise 3 Recent research suggests that Japanese women and men may use grammatical patterns with different frequencies. Are you aware of any differences in the grammar of English-speaking women and men? What pattern of gender differences would you predict for grammatical variables such as multiple negation, which was discussed in chapter 6? Answer at end of chapter Across all social groups in Western societies, women generally use more standard grammat- ical forms than men and so, correspondingly, men use more vernacular forms than women. In Detroit, for instance, multiple negation (e.g. I don’t know nothing about it), a vernacular feature of speech, is more frequent in men’s speech than in women’s. This is true in every social group, but the difference is most dramatic in the second highest (the lower middle class) where the men’s multiple negation score is 32 per cent compared to only 1 per cent for women. Even in the lowest social group, however, men use a third more instances of multiple negation than women (90 vs 59 per cent). This pattern is typical for many grammatical features. In many speech communities, when women use more of a linguistic form than men, it is generally the standard form – the overtly prestigious form – that women favour. When men use a form more often than women, it is usually a vernacular form, one which is not admired overtly by the society as a whole, and which is not cited as the ‘correct’ form. This pattern has been found in Western speech communities all over the world. It was described in 1983 by Peter Trudgill, the socio- linguist who collected the Norwich data, as ‘the single most consistent finding to emerge from sociolinguistic studies over the past 20 years’. This widespread pattern is also evident from a very young age. It was first identified over thirty years ago in a study of American children’s speech in a semi-rural New England village, where it was found that the boys used more [in] and the girls more [ih] forms. Later studies in Boston and Detroit identified the same pattern. Boys used more vernacular forms such as consonant cluster simplification: e.g. las’ [las] and tol’ [toul], rather than standard last [last] and told [tould]. Boys pronounced th [e] in words like the and then as [d] more often than girls did. In Edinburgh, differences of this sort were observed in the pronunciation of girls and boys as young as 6 years old. The pattern is clear, consistent and widespread and it is evident from a very early age. What is the explanation for it? Why does female and male speech differ in this way? Exercise 4 Before you read the next section, consider some possible explanations for the finding of urban social dialect surveys that women use more standard forms than men. Consider the possible influence of the dimensions discussed in chapter 1: social status, social distance or solidarity, the formality of the context and the functions of speech. How might these affect the speech used by an interviewee in a social dialect survey? Bear in mind that no single explanation is likely to fit all cases. 166  Chapter 7 Gender and age Explanations of women’s linguistic behaviour ‘Why can’t a woman be more like a man?’ (My Fair Lady) When this pattern first emerged, social dialectologists asked: ‘why do women use more standard forms than men?’ At least four different (though not mutually exclusive) explana- tions were suggested. The first appeals to social class and its related status for an explanation, the second refers to women’s role in society, the third to women’s status as a subordinate group, and the fourth to the function of speech in expressing gender identity, and especially masculinity. The social status explanation Some linguists have suggested that women use more standard speech forms than men because they are more status-conscious than men. The claim is that women are more aware of the fact that the way they speak signals their social class background or social status in the community. Standard speech forms are generally associated with high social status, and so, according to this explanation, women use more standard speech forms as a way of claiming such status. It has been suggested that this is especially true for women who do not have paid employment, since they cannot use their occupations as a basis for signalling social status. The fact that women interviewed in New York and in Norwich reported that they used more standard forms than they actually did has also been used to support this explanation. Women generally lack status in the society, and so, it is suggested, some try to acquire it by using standard speech forms, and by reporting that they use even more of these forms than they actually do. Though it sounds superficially plausible, there is at least some indirect evidence which throws doubt on this as the main explanation for gender differences in social dialect data. It is suggested that women who are not in paid employment are most likely to claim high social status by using more standard forms. This implies that women in the paid workforce should use fewer standard forms than women working in the home. But the little evidence that we have in fact suggests that just the opposite may be true. An American study compared the speech of women in service occupations, working in garages and hotels, for instance, with the speech of women working in the home. Those in paid employment used more standard forms than those working in the home. In the course of their jobs, the first group of women were interacting with people who used more standard forms, and this interaction had its effect on their own usage. By contrast, the women who stayed home interacted mainly with each other, and this reinforced their preference for vernacular forms. Exactly the same pattern was found in an Irish working-class community. The younger women in Ballymacarrett, a suburb of Belfast, found work outside the community, and used a much higher percentage of linguistic features associated with high status groups than the older women who were working at home. This evidence throws some doubt, then, on suggestions that women without paid employment are more likely to use standard forms than those with jobs, and so indirectly questions the social status explanation for women’s speech patterns. A variation on this explanation suggests that standard or prestige forms represent linguistic capital which people can use to increase their value or marketability in some contexts. This has the advantage of accounting for the higher proportion of such forms in the speech of those in the white collar professional workforce, especially when they are interacting with 167 An introduction to sociolinguistics people they want to impress. Where women have few other sources of prestige, language may become especially significant as a social resource for constructing a professional identity. But if you work in a soap factory or a shoe factory, or on a building site, the forms that your companions value are more likely to be vernacular fo