Lessons from a survey of British dialect grammar

Links & Letters 5, 1998 61-73 Abstract This article is intended for non-specialists interested in linguistic and educational issues associated with regional variation in grammar. It is based on the work of the Survey of British Dialect Grammar which entailed an unusual collaboration between linguists and school children. The findings of the survey have made it possible to provide a fuller picture of the national distribution of nonstandard grammatical features and to generate hypotheses about dialect levelling in urban centres. They also provide interesting challenges to educational policy and practice. Key words: British Dialects, Non-standard English, Sociolinguistics, Education. Table of Contents 1. Introduction Regional variation in English has long been the object of scholarly attention. With the emergence of sociolinguistics as an independent discipline, scholarly interest in this question has continued to flourish (Trudgill 1990; Trudgill and Chambers 1991; Milroy and Milroy 1992). However, coverage of the different parts of the British Isles remains extremely patchy with individual linguists tending to focus on their home territories. For many areas, popular publications such as Notts natter: how it is spoke (Wright 1986) and Son of Bristle: a second guide to what the natives say and mean in the heart of Wess Vinglun (Robson 1982) are the main sources of information. (*) The research reported in this paper was supported by ESRC research award no. C-00-23-2264. 1. Introduction 2. The survey of British dialect grammar 3. Some unexpected outcomes 4. Resources for diversity 5. Conclusion References Lessons from a survey of British dialect grammar * Jenny Cheshire Queen Mary and Westfield College, University of London Viv Edwards Faculty of Education, University of Reading September 1997 62 Links & Letters 5, 1998 Jenny Cheshire; Viv Edwards To date there have been two national surveys of British dialects. The first, the Survey of English Dialects, grew out of the work of Harold Orton and his colleagues (1962–71) at the University of Leeds. Pre-dating as it does the advent of modern sociolinguistics, the focus is on lexis and phonology. The second national overview of regional variation, the Survey of British Dialect Grammar, was made possible by an Economic and Social Research Council Research award to Jenny Cheshire and Viv Edwards. As the title implies, there was a significant shift from lexis and phonology to grammar. Other changes in focus concerned methods of data collection. Traditional dialectologists elicited information from elderly, male informants. In contrast the more recent survey focuses on school children in mainly city schools. A partnership between researchers and schools came into being for very practical reasons. The project’s shoestring budget was not sufficient to employ trained linguists to record and analyze authentic speech data in a wide range of settings. However, at the time the project was taking place, there was considerable interest in regional and social variation in schools. By providing stimulus material to teachers, it was possible to harness this interest in order to collect a wide range of interesting data, using school children both as informants and as researchers. We were also able to collect qualitative data on language attitudes from older informants: coverage of the survey in regional newspapers and local radio broadcasts resulted in many letters containing personal views on dialect. 2. The survey of British dialect grammar In the first stage of the project, we established a network of teachers throughout the United Kingdom. Some 87 schools distributed throughout the country ultimately took part in the survey. While the coverage of rural areas was incomplete, urban areas were well-represented. Approximately 5,000 children took part. The number of participants varied from school to school: in some cases, teachers worked with one class, in other cases, they worked with several. Almost all the schools were comprehensive schools drawing on children from a wide range of socio-economic backgrounds and pupils between the ages of eleven and sixteen. For reasons of economy, the survey took the form of a questionnaire, which we sent to all the participating schools. The questionnnaire consisted of 196 linguistic features, drawn from the main areas of dialect grammar described in Edwards, Trudgill and Weltens (1984). We felt it essential that a period of preparatory work should precede the administration of the questionnaire, in order to ensure that children provided reliable information, rather than the answers they assumed their teachers wanted. In order to reinforce this point, a series of lesson outlines and materials were sent to all participating schools. The lesson outlines covered topics such as multilingual Britain, language variation, language change, standard English, and ‘talking proper’. The questionnaire on local dialect usage was presented as the end Lessons from a survey of British dialect grammar Links & Letters 5, 1998 63 point of the work around social and regional variation in English, with the intention of consulting pupils as the experts on their local variety of English, and asking them to tell us whether the forms listed on the questionnaire were used locally. We suggested to teachers that classes working collaboratively should divide into three groups, each dealing with one page of the questionnaire, and that each group should report on the forms of dialect grammar listed on their page that were used in their community. If more funds and research staff had been available it would, of course, have been preferable to have based this part of the survey on audio recordings of a sample of speakers in different parts of Britain. We were, of course, acutely aware of the limitations of using questionnaires to collect linguistic data and we took steps to guard against these limitations, as far as possible. A pilot study was carried out in 1986 in the town of Reading, Berkshire, where a previous empirical study of morphological and syntactic variation had taken place (Cheshire 1982). In the pilot study, children’s appraisals of dialect forms regularly heard in Reading coincided closely with those which we know, from the previous study, actually do occur. Each completed questionnaire that we received during the survey itself was examined to see if any examples of dialect usage were reported which, on the basis of existing knowledge, were unexpected for the area. Such examples were infrequent but did occasionally occur. Our procedure was to write to the teacher concerned to query the feature and to ask for further examples of utterances in which it occured. This allowed us to judge for ourselves whether misreporting had taken place. Wherever possible, we also cross-checked examples of this kind with linguists working in that area of Britain. Several problems emerged during the data collection phase. Teachers were not able to cooperate as fully as we had hoped partly because of industrial action which they were taking on pay and conditions during this period and partly because of the pressures caused by the introduction of a new examination system. Two hundred questionnaires were distributed, and eighty-seven returned. Although this was a smaller number than we originally expected, the completed questionnaires in fact covered all the major urban areas of England, Scotland and Wales with the exception of Edinburgh, Newport and Portsmouth (see Cheshire, Edwards and Whittle 1989). The Survey of British Dialect Grammar therefore contrasts sharply with the Survey of English Dialects not only in its focus on syntax rather than on phonology, but also in its emphasis on urban rather than rural areas. This predominantly urban distribution of responses allowed us to make a contribution to the controversial question of dialect levelling. Earlier analyses of English dialects based on the Survey of English Dialects material (see, for instance, Wakelin 1984; Lass 1987) found it possible to define regional dialect areas in terms of the phonological features that occur in different parts of Britain, but could not identify any clearcut regions of England in terms of morphological features. Lass (1987: 234) suggests that this may be because 64 Links & Letters 5, 1998 Jenny Cheshire; Viv Edwards morphosyntax has remained relatively stable throughout its history; Hudson (1983) also draws attention to the fact that, on the whole, there is less syntactic than phonological variation in language. The massive social and demographic changes that have taken place since the Second World War, however, appear to have had an effect on this putative stability. Urban dialectologists agree that the growth of cities has been accompanied by very rapid mixing of a number of different dialects from surrounding areas (see Milroy 1984: 214), as former rural populations become increasingly urbanized. Dialects of English are now usually thought of as falling into two groups: traditional dialects, spoken by a probably shrinking minority of speakers living mainly in remote and long-settled rural communities, and mainstream dialects, spoken in various parts of the English-speaking world, including most of the urban areas of Britain (see Trudgill and Chambers 1991: 2-3). Whilst traditional dialects may differ from each other and from standard English in unpredicable ways, mainstream urban dialects are thought to closely resemble one another and to have relatively few grammatical differences from standard English. In other words, it is thought that in some cases dialect diversity is reducing and being replaced not simply by standard grammatical forms but also by a development towards a levelled nonstandard dialect. This is a controversial question that can only be properly addressed by empirical investigations of actual usage, but the survey responses allowed us to make a preliminary, informed contribution to the question, by determining those features that were reported most frequently as used in the urban centres of Britain. Certain nonstandard grammatical features, such as multiple negation and unmarked plurality, are sometimes listed as common to most urban varieties of English (see, for example, Hughes and Trudgill 1987). Coupland (1988: 35) suggests that several of these grammatical features are so widespread that they are best seen as British social dialect characteristics rather than as marking regional provenance. These assumptions, however, have been unsupported by systematically collected empirical data. The survey responses formed the basis for a more principled approach than has previously been possible. It emerged that a large number of features were reported infrequently (61 of the 196 features —31 per cent— were reported by fewer than 5 per cent of the schools). One dialect feature (demonstrative them, as in them big spiders ), on the other hand, was reported by more than 90 per cent of the schools who took part in the survey, and a further ten features were reported by more than 80 per cent of the schools. We list these eleven features below, together with the percentage frequency with which they were reported and the questionnaire item that was used to ask about them. Note that we attach no importance to the actual percentage frequencies; these were calculated simply as a way of distinguishing those features that were reported more widely than others. Lessons from a survey of British dialect grammar Links & Letters 5, 1998 65 them as demonstrative adjective (item 125: Look at them big spiders) 97.7 % should of (item 196: you should of left half an hour ago!) 92.0% never as past tense negator (item 7: No, I never broke that) 86.2% absence of plural marking (item 95: To make a big cake you need two pound of flour) 86.2% what as relative pronoun (item 115: The film what was on last night was good) 86.2% there was with plural ‘notional’ subject (item 58: there was some singers here a minute ago) 85.1% there’s with plural ‘notional’ subject (item 29: There’s cars outside the church) 83.9% present participle sat (item 46: She was sat over there looking at her car) 83.9% nonstandard was (item 51: we was singing) 83.9% adverbial quick (item 86: I like pasta. It cooks really quick) 82.8% ain't / in't (items 9 and 10: that ain't working / that in't working) 82.8% present participle stood (item 47: And he was stood in the corner looking at it) 80.5% The survey thus confirms that the following features are widespread throughout the urban centres of Britain, as suggested by Hughes and Trudgill (1987) and Coupland (1988): them as demonstrative adjective, absence of plural marking on nouns of measurement, what as relative pronoun, nonstandard was, adverbials without the - ly suffix, and ain’t/in’t. In other words, ‘regional dialect’ appears to be a misnomer for these features. We think it important to report, however, that their social distribution and frequency of use has yet to be determined, so that, despite Coupland’s (1988) suggestion, we do not yet feel ready to label any of them as social dialect features. Some features were not reported as frequently as we had expected. These include multiple negation; the use of simple prepositions such as up, round and over where standard English has complex prepositions such as up to, round to or over at; the regularized reflexive pronoun forms hisself and theirselves; done as the past tense form of full verb do . With the exception of the nonstandard reflexive pronoun forms, all these features were reported much more frequently by schools in the south of England than by schools else- 66 Links & Letters 5, 1998 Jenny Cheshire; Viv Edwards where in Britain, suggesting that there may be a hitherto unsuspected regional distribution to these forms. 3. Some unexpected outcomes The results of the survey provided the information that we had hoped for, giving us a general picture of the regional distribution of those features of dialect grammar included in the questionnaire at the same time as generating some preliminary hypotheses about dialect levelling in the British Isles. In addition, we obtained some sociolinguistic information of a more general kind, as a by-product of the lesson suggestions. Children’s written work was often of considerable interest to us as sociolinguists because it provided a qualitative counterbalance to experimental research on attitudes to regional variation in English. We give some examples in the following sections. 3.1. Reactions to linguistic diversity We were particularly interested in the topic of correction, both because we wanted to see which linguistic variables are salient to the teachers, parents and other adults who feel the need to monitor children’s language and because we were curious about the children's reactions. We were also interested in written work that illustrated the way in which language functions as a symbol of individual and social identity, or which revealed children’s attitudes to regional variation. 3.1.1. Correction Part of the lesson outline on the topic of ‘talking proper’ invited children to reflect on correction. We were interested in the identity of the forms that pupils said were corrected, since there is little precise information available on stigmatized forms in the research literature, and few hypotheses offered to explain why some variables are more salient to speakers than others. Condemnation of ‘h’ dropping, for example, is very widespread in Britain, but people do not seem to be concerned about vowel alternations (such as regional variation in the pronunciation of a word such as bus , which may be [b à s] or [b υ s]). Trudgill (1986: 11) suggests that overt stigmatization occurs when there is a high status variant of the stigmatized form which tallies with the orthography, while the stigmatized form does not. Many of the corrections concerning pronunciation that schoolchildren wrote about gave support to this view. Sometimes (as in examples 1-3) corrections were expressed in terms of ‘dropped letters’ or additional ‘letters’: (1) Mum corrects my speech when I drop letters especially ‘h’ and it annoys me but I suppose she’s right. Lessons from a survey of British dialect grammar Links & Letters 5, 1998 67 (2) Yes they moan at me when I start to speak like a Scouser [someone from Liverpool]. I say married as if there’s about seven r’ s in it - marrrrrrried. (3) Yes like when I say ye they always correct me and say yes . Predictably, other corrections concerned features of nonstandard English grammar, which —it was claimed— did not exist (example 4) or were ‘not English’ (example 5): (4) Yes because I use words like worser and other things like that when there’s no such word. (5) Yes they correct me when I am saying something and say that’s not English. If I say what they say that . If I say can I borrow this they say it’s not borrow it’s lend . A second reason for our interest was the information that we received, indirectly, on whether correcting children’s language is a worthwhile exercise. Teachers in England and Wales are sometimes advised on the kinds of corrections that they should make. For example, the Cox Report (DES 1989: 4.46) rightly points out the dangers of indiscriminate correction, but suggests that teachers should correct nonstandard forms and highly stigmatised forms (such as the past tense forms of see ) that occur frequently. The work which we received, however, led us to believe that such an approach is often counterproductive. Several children who reported that their speech had been ‘corrected’ gave examples which could not be attributed to standard English speaking teachers and can be explained more adequately in terms of hypercorrection: (6) Teachers normally correct me when I say can I lend a pen but you should say can you please borrow me a pen . (7) Yes. When I say I saw something they [teachers] say to say seen but my parents say it the opposite. This confuses me. It can be argued on the basis of these examples that correcting pupil’s speech is a waste of time and is likely to lead to confusion about the linguistic relationship between features of standard and nonstandard English. Sociolinguists stress that language is closely bound up with individual and social identity: criticisms of the way we speak are likely to be interpreted as personal attacks. It is certainly not dificult to envisage a scenario in which persistent corrections of a child’s language can lead to a reticence in oral work. We were interested, therefore, in the reactions that children expressed to being corrected. Examples 8-11 illustrate a broad range of professed reactions: (8) I feel very angry because I know what I am saying and so does the teacher. (9) I am not really bothered: I know what I mean and so do they. 68 Links & Letters 5, 1998 Jenny Cheshire; Viv Edwards (10) My mum corrects me and it annoys me but I suppose she’s right. (11) It doesn't bother me because they [teachers] know how to speak better than I do. With the possible exception of the sanguine response in example 9, this range of reactions confirms our view that it is not a good idea to correct children's speech. In this respect, the repeated complaints of older dialect speakers also consulted during the course of the survey present a point of comparison. One northern octogenarian commented: Any child using dialect speech would be severely reprimanded or ignored, depending on which teacher was in charge. Some teachers would endeavour kindly to explain that this was not on; others, less sympathetic, would perhaps resort to sarcasm or pretend to deliberately misunderstand. Although tolerance of the spoken word has certainly increased in recent years, it is extremely doubtful whether the attempts of present-day teachers to change dialect speech or writing will be any more effective than those of the past. As one fourth year pupil in Rotherham reflected: Teachers always correct the way I write. They correct the way I write more than anything. When I write a story and include talking I write it how I would speak. But sometimes teachers cross it out and put in how they would talk. I don’t think they should do that. They should leave it as it is. Though we remain sceptical about the usefulness of correcting nonstandard forms, we are convinced of the value of including discussion of this aspect of ‘talking proper’ in the school curriculum. We agree with the Cox Committee’s view that every child is entitled to learn not only the functions but also the forms of standard English (DES 1989: 4.7). However, it is by no means clear how this should be achieved, nor how teachers are to achieve the Cox Report’s instructions to teach children the grammatical differences between the speech of their local area and spoken standard English (DES 1989: 15.37ii). Nonetheless, the opportunity to air personal reactions to corrections and to see the range of reactions amongst classmates seems a necessary prerequisite for constructive teaching of the linguistic differences between standard and nonstandard English. 3.1.2. Attitudes to regional variation The research literature in sociolinguistics and social psychology is unanimous about the nature of attitudes towards regional varieties of English. Experiments have shown that accents associated with rural areas of Britain tend to be perceived by British speakers of English as more attractive than accents spoken in heavily urbanised areas, and that speakers with Received Pronunciation Lessons from a survey of British dialect grammar Links & Letters 5, 1998 69 (sometimes called ‘the Queen's English’) are considered to be more intelligent and more competent than speakers who have a regional accent (Edwards 1990). This perception has been found to be shared by standard and nonstandard speakers alike, though nonstandard speakers may have strong feelings about the value of their own speech, associating it with friends, family and neighbourhood, with social attractiveness and with integrity. The lesson suggestions invited children to consider their views on regional dialects. One teacher in a school in the urban centre of Widnes, Lancashire (in north west England) sent us some written work that showed that ‘talking posh’ was often associated with the south of England, particularly London: (12) I dislike London accent. It sounds really posh. (13) I dislike London accent because they are stuck up snobs. As for other accents, the most striking feature of the children’s comments was the complete lack of unanimity in their likes and dislikes, and the very wide range of reasons that were given in support of these opinions. Many of the attitudes that were expressed within a single class of pupils directly conflicted with each other. Compare, for example, 15 with 16, 17 and 18, 19 with 20, and 21 and 22 with 23 and 24: (14) I like Cockney because it gives you a laugh. I also like American because it's dead cool. Geordie is OK as well. (15) I dislike Geordie accent because of the way they say it, it just gets right up my nose. (16) I detest American accents because there is too much of it going on TV. (17) I don't like Cockney accents because it sounds like they’re talking out of their nose (18) I like the Welsh, Manchester, and Australian accents because they're good. (19) I dislike Manchester accent because I don't like Manchester and everything's slower. (20) I like Scottish because of the way they say it and when they say it fast it sounds dead cool. (21) Scottish is the best because it sounds so easy going. (22) I dislike Scottish because I can't understand what they are saying. (23) I dislike Scottish accent because they speak so quick I can't understand it. The comments that pupils made about different regional accents revealed a very interesting selection of idiosyncratic likes and dislikes, which they justified in equally idiosyncratic ways: (24) Norfolk accent is the best. The people sound like farmers.