Adverbs and social class 1 Ronald Macaulay Pitzer College, California An earlier study, based on interviews with a socially stratied sample showed a dierence in the use of adverbs, with the middle-class speakers using derived adverbs in -ly more than twice as frequently as the working-class speakers. An examination of interactions in peer-group same-sex dyads shows a similar socially stratied pattern in both adults and adolescents. There are similar dierences in the use of some other adverbs and certain adjectives. The consistency of the results suggests that there is a stable dierence in speech styles between the two social classes and that this dierence reects a dierent attitude on the part of the speakers to their audience. KEYWORDS: Social class, discourse analysis, speech style, Scottish dialect, age dierences I am glad you like adverbs I adore them; they are the only qualications I really much respect . . . Henry James, letter to a young admirer, 1902 INTRODUCTION Social class gured prominently in early sociolinguistic investigations (Macau- lay and Trevelyan 1973; Trudgill 1974) but became less central as the focus moved to ethnicity, networking (Milroy 1980) and gender (Coates 1986). There continue to be studies of social class dierences (e.g. Foulkes and Docherty 1999) usually with particular attention to the role of social class dierences in language change. There have, however, been fewer studies of stable social class dierences. Macaulay (1991) examined a range of features in a small sample of speakers in the town of Ayr in the west of Scotland. Many of the clearest social class dierences were in pronunciation and morphology, similar to those found in other sociolinguistic studies, but there were also dierences in syntax, lexical choice and discourse features. Some of the latter dierences resembled those # Published by Blackwell Publishers Ltd. 2002 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden MA 02148, USA. Journal of Sociolinguistics 6/3, 2002: 398417 ADVERBS AND SOCIAL CLASS 399 # Blackwell Publishers Ltd. 2002 examined by Basil Bernstein in his early attempts to characterize social class dierences in language (Bernstein 1962). As a young man working with boys' clubs in the east end of London and later teaching adolescents at a day college, Bernstein was struck by the dierence between the boys' verbal skills and their performance skills (Bernstein 1971: 2 5). He later demonstrated this by comparing the results of two polarized groups on tests of verbal and non-verbal intelligence. He was able to show that the verbal scores of the working-class boys were depressed in comparison with their non-verbal scores, while there was no dierence for the middle-class subjects (Bernstein 1960). Bernstein had some intuitions about the actual linguistic dierences that he set out in a paper not based on his own research (Bernstein 1959). Among the characteristics of what he was then calling a public language (used by, among others, `the unskilled and semi-skilled strata') was `rigid and limited use of adjectives and adverbs' (1971: 42). 2 Bernstein later investigated this notion empirically in a study based on discussion groups with two middle- class groups of ve boys and two working class groups of ve boys and one of four (Bernstein 1962). A sample of 1,800 words was taken from each session but this number was reduced by the omission of certain forms and the contribution of two working-class boys who contributed fewer than 90 words each. The total number of words analyzed was 7,892. Bernstein identied a category of `uncommon adverbs' by excluding adverbs of degree and place, just, and really, in addition to a number of items that would not normally be considered adverbs (e.g. not, how). Bernstein does not give the actual frequencies but only the results of the statistical analysis: `[a] greater proportion of the adverbs of the middle class are uncommon and the dierence is signicant beyond the 0.001 level of condence' (1971: 101). This dierence in adverb use has similarities to one found in the Ayr study, and I later examined a corpus of conversations recorded in Glasgow (Stuart-Smith 1999) where the same social class dierence in the frequency with which the speakers used adverbs emerged. This paper reports the results from the two studies and explores possible explanations for this consistent dierence. THE SAMPLE In 1978 and 1979 I conducted interviews in Ayr as part of a proposed (but never completed) comparative study of urban speech in Scotland. From these interviews I chose twelve speakers, six middle-class and six lower-class, for detailed analysis (Macaulay 1991). The sample was clearly polarized in social class terms on grounds of occupation, education and residence. The tapes were transcribed in their entirety and searched for phonological, morphological, syntactic, lexical and discourse features (Macaulay 1991). The size of the corpus is 120,669 words (lower-class speakers 69,711; middle-class speakers 50,898). The second set of recordings was collected for an investigation of language variationand change inGlasgow, Scotland (Stuart-Smith1999). The study is one MACAULAY 400 # Blackwell Publishers Ltd. 2002 of several (Foulkes and Docherty 1999) carried out to discover what changes, if any, had occurred in British urban speech since the earlier studies of the 1970s (e.g. Macaulay and Trevelyan 1973; Trudgill 1974). In the summer of 1997, 33 Glaswegians were recorded in same-sex dyadic conversations of approximately 35 minutes long. The speakers were drawn from two areas of the city, representing broadly urban working-class and suburban middle-class areas. The sample consists of two age groups: adolescents (1314) and adults (40+), with equal numbers of males and females. 3 For each session one speaker was selected and asked to choose someone they would feel comfortable talking to in the presence of a tape-recorder for about half an hour. The participants were free to talk about anything they wished. The resulting tapes provide material for an examination of age, social class and gender dierences in this particular form of discourse, and are free from any addressee eect (Bell 1984) that might be caused by an academic interviewer. Although there are more adult speakers in the Glasgow study the total amount of speech recorded from the adults is less than in the Ayr study, 84,616 words (working-class speakers 50,307; middle- class speakers 34,309). The main reason is that the Ayr interviews lasted longer than half an hour. Inthe Glasgowsample, unlike the Ayr sample, there were also adolescents aged 1314. They produced considerably less speech overall than the adults, 43,046 words (working-class adolescents 21,093; middle-class adolescents 21,953) and there were social class and gender dierences in the amount of talk recorded. The working-class boys produced less speech than the others and the working-class girls the most, so the adolescent corpus is unbalanced in gender terms but there is enough speech from all four categories to justify quantitative analysis. Bernstein emphasized that the kinds of dierences he found between working- class and middle-class speakers were not absolute but relative: `[t]he dierence on individual measures was always one of relative frequency' (1971: 13). The key measure used in the present paper is the frequency of use per one thousand words. The Ayr and Glasgow samples were transcribed in their entirety, both as dialogues and with the contribution of each speaker separated. The contributions of individual speakers were analyzed by means of the WordCruncher concord- ance programme. The resulting lists were then manually searched for items that might vary in their distribution. Among these were derived adverbs in -ly. THE DATA 1. Adverbs in -ly Adults. Table 1 gives the frequency of adverbs in -ly for the Ayr sample, from Macaulay (1995: 44), and for the Glasgow adult sample. It can be seen that while there are minor dierences, the general pattern is remarkably similar in both, with the middle-class speakers using derived adverbs in -ly more than twice as frequently as the working-class speakers. This similarity was reassuring ADVERBS AND SOCIAL CLASS 401 # Blackwell Publishers Ltd. 2002 because I had worried that the social class dierences in the use of derivative adverbs found in the Ayr study might be an artifact of the interview situation (Macaulay 1995). The gures in Table 1 cannot be directly compared with Bernstein's ndings because they include really and degree adverbs, and Bernstein reports only the proportion of `uncommon adverbs' rather than frequencies, but the pattern is presumably similar. Figure 1shows that for the Glasgowadults the individual frequencies reect the Table 1: Relative frequency of derivative adverbs in -ly in Ayr and Glasgow Ayr Glasgow Lower-class Middle-class Working-class Middle-class # Freq. # Freq. # Freq. # Freq. Manner 28 0.40 82 1.61 11 0.22 32 0.93 Time/Freq. 41 0.58 70 1.38 19 0.38 33 0.96 Degree 47 0.67 121 2.38 35 0.69 42 1.22 Sentence 76 1.08 174 3.42 92 1.82 197 5.74 really 55 0.79 106 2.08 93 1.85 104 3.03 Totals 247 3.52 553 10.87 250 4.97 408 11.89 Freq. = per 1,000 words Figure 1: Frequency of use of adverbs in -ly by 18 Glasgow adults (freq. = per 1,000 words) MACAULAY 402 # Blackwell Publishers Ltd. 2002 general pattern with two outliers, one middle-class woman with a frequency of 6.5 and one working-class woman with a frequency of 12.2. Figure 1 also shows that it is the middle-class men who are the most frequent users of these adverbs. At this point it may be helpful to point out that while the dierences in adverb use are not salient and are not indexical of social class membership in Silverstein's (1996) sense, there are many obvious dierences in pronunciation and morphology that distinguish the two groups. The dierences between the two groups in Ayr are summarized in Macaulay (1991: 257). The dierences in pronunciation in Glasgow are presented in Stuart-Smith (1999). Nobody from Ayr or Glasgow would have the slightest diculty in assigning any of the speakers to one social class or the other on the basis of a short extract from the tapes. The two groups are clearly polarized within the local speech community. Macaulay (1995) also examined the use of adjectives by the two groups of speakers in Ayr and found that the middle-class speakers used adjectives with a frequency of 22.41 per thousand in contrast to the lower-class speakers with a frequency of 11.74. The gures for the Glasgow sample show the middle-class speakers with a frequency of 34.16 and the working-class speakers with a frequency of 24.74. Once again, the pattern is repeated, though the distance between the groups is less in Glasgow. Adolescents. What about the adolescents in Glasgow? The overall frequency of derivative adverbs in -ly for the Glasgow adolescents is given in Table 2. Although the overall frequency of use is lower than in the adult sessions the pattern of social class dierences is similar and the dierence is still substantial. The individual gures are shown in Figure 2. Here the outliers are two middle- class boys who use very few adverbs in -ly and one working-class boy who uses these adverbs with a frequency of 8.2. It is the other two middle-class boys that use these adverbs most frequently, like the middle-class men. Given the overall higher frequency with which these adverbs are used in the middle-class sessions, it is not surprising that the variety of adverbs is greater. In Table 2: Frequency of adverbs in -ly in Glasgow adolescent conversations, by gender and social class N Freq. Middle-class girls 84 8.1 Middle-class boys 101 8.8 Working-class girls 50 3.7 Working-class boys 29 3.9 MC adolescents 185 8.4 WC adolescents 79 3.9 Freq. = per 1,000 words ADVERBS AND SOCIAL CLASS 403 # Blackwell Publishers Ltd. 2002 the middle-class adult conversations there are 74 dierent adverbs in -ly used; in the working-class conversations, 37 dierent adverbs are used. Of the total, 22 adverbs are used by both groups, the most frequent being really (MC 3.03/ WC 1.85) and actually (MC 2.8/WC 0.74). In the case of the adolescents, the middle-class speakers use 32 dierent adverbs in -ly, and the working-class 21, with 15 used by both. As with the adults, the most frequent are really (MC 3.14/WC 0.85) and actually (MC 0.73/WC 0.57). However, the list of adverbs unique to either social class group does not suggest that the source of the dierence lies in education. It is unlikely that educational dierences account for the failure of working-class speakers to use adverbs such as badly, clearly, fairly, happily, etc. or their use of such adverbs as automatically, basically, entirely, and literally that do not occur in the middle-class conversations. Nor is there any indication that the working-class speakers are using uninected forms (suxless adjectives, zero forms) instead of inected adverbs. These forms are more common in American English (Opdahl 2000) and they have also been found to occur fairly frequently in northern British English (Tagliamonte and Ito 2001). In both the Ayr and the Glasgow recordings uninected forms were too rare to aect the results. As regards the use of adjectives, the social class dierence found among the adults is repeated in the adolescent conversations. The middle-class adolescents use adjectives with a frequency of 29.79 per thousand words and for the working-class adolescents the frequency is 21.86. The social class dierences Figure 2: Frequency of use of adverbs in -ly by 16 Glasgow adolescents (freq. = per 1,000 words) MACAULAY 404 # Blackwell Publishers Ltd. 2002 that were found in the Ayr interviews have thus been repeated in the Glasgow conversations, among both adults and adolescents. 2. Other adverbs It is not only derived adverbs that show this marked social class dierence. In the Ayr interviews the middle-class speakers used very with a frequency of 3.03 per thousand words compared with a frequency of only 0.82 in the lower-class interviews. A similar pattern is found in the Glasgow conversations (see Figure 3). It can be seen from Figure 3 that very is almost categorically a middle-class word. Eleven (61%) of the 18 working-class speakers do not use it even once in their conversations. The Glasgow adolescents have two additional intensiers that are not found in the adult sessions, pure and dead. The frequencies are shown in Figure 4. Examples of the use of pure and dead can be seen in 1: 1. a. this is pure embarrassing b. this is dead embarrassing c. it's pure funny but d. I'd look dead funny without a fringe wouldn't I? e. and I was like really close to Chi f. I was standing pure close to him g. she used to be dead fat h. she's dead skinny now Examples 1e and 1f show that pure is most likely an alternative to really but many of the examples of dead are like those in 1g and 1h modifying an adjective Figure 3: Frequency of use of very and quite by Glasgow adults and adolescents (freq. = per 1,000 words) ADVERBS AND SOCIAL CLASS 405 # Blackwell Publishers Ltd. 2002 and could be an alternative to very. It can be seen from Figure 4 that while both boys and girls use these forms, girls use themthree times as frequently. The overall social class dierence for the two intensiers is very slight (MC 5.4/WC 6.4). Another adverb whose use varies socially is quite. In the Ayr interviews, the middle-class speakers also used quite more frequently (2.49 per 1,000 words) compared with the lower-class speakers (1.00 per 1,000 words). A similar pattern can be seen in the Glasgow conversations, as shown in Figure 3, which shows that quite is predominantly a middle-class item in both age-groups. The use of quite can either be emphatic (what Quirk, Greenbaum, Leech and Svartvik 1985: 590 call `maximizers'), as in the examples in 2, or a hedge (what Quirk et al. 1985: 577578 call `downtoners'), as in the examples in 3. All the examples in 2 and 3 are from middle-class conversations: 2. a. but I think clothes-wise we're quite dierent b. and I was quite proud because I was still thirty nine c. I do it quite quickly I can do it in about fteen seconds d. San Francisco's actually quite chilly so e. it's quite quite quite quite dierent 3. a. it is actually quite nice b. I mean Alison's still quite sort of young c. the actual wee beach is is quite nice because it's sort of rough sand d. it's quite pleasant it's it's sand-dunish and em e. but it's it's er quite interesting to nd how dierent people do speak Figure 4: Frequency of use of pure and dead by Glasgow adolescents (freq. = per 1,000 words) MACAULAY 406 # Blackwell Publishers Ltd. 2002 Deciding between these two functions is sometimes dicult so any gures reect an interpretative decision. The middle-class speakers appear to use quite more frequently in its emphatic function (67%) than in its hedging function (33%). For the working-class speakers the dierence is less: 56 percent emphatic, 44 percent hedging. However, the middle-class speakers use quite with an overall frequency of 3.64 per thousand words compared with the working-class frequency of 1.19. The frequency with which the middle-class speakers use quite in its emphatic function is 2.42 per thousand words compared with the working-class frequency of 0.66. In the hedging function the frequencies are: middle-class 1.2, working-class 0.52. The middle-class thus use quite twice as often as the working-class speakers in a hedging function and almost four times as often in the emphatic function. The nal adverb to be examined is just. In the Ayr interviews the dierence in the frequencies of just was minimal (MC 5.01/WC 4.84). The frequencies for the Glasgow speakers are shown in Figure 5. It can be seen that the social class dierences are slight. This is the only example of a very common adverb that the working-class adults use more frequently (6.72 per 1,000 words) than the middle-class adults (5.28). There are, however, also social class dierences in the use of just. While the working-class adults use just slightly more frequently (6.18 per 1,000 words) than the middle-class adults (5.22 per 1,000 words), they do not use just in exactly the same way. In Ayr, following the analysis presented in Lee (1987), I separated the uses of just into four categories. The rst is with reference to time, usually the Figure 5: Frequency of use of just by Glasgow adults and adolescents (freq. = per 1,000 words) ADVERBS AND SOCIAL CLASS 407 # Blackwell Publishers Ltd. 2002 immediate past, as in examples 1a and 1b. The second use is as an intensier with the general sense of `exactly,' as in examples 2a and 2b. The third use is in the sense of `only,' as in examples 3a and 3b. Finally, there is the sense of `simply,' as shown in examples 4a and 4b: 4. (the a examples are middle-class, b examples working-class) 1a. I've just realised something (speaker 10R) 1b. that's it just opened up again (13R) 2a. yes that's just what I was thinking (12L) 2b. well just as it turns round the bend (15L) 3a. but it's just a baby (12L) 3b. it was just the two of us (14L) 4a. I'll just take everything out of the dining room (10R) 4b. I'll just go alang (13R) The examples in 4 show that both groups use just in all four senses but they do not use them equally frequently as shown in Figure 6. The working-class adults use just more often in the `simply' sense, while the middle-class adults make more frequent use of the `recency' and `exactly' senses than do the working- class speakers. The latter middle-class use is most distinctive when employed emphatically with adjectives and verbs as in 5: 5. a. It's just awful. I mean that's my lot plus another three it's just horrendous you know absolute madness (10R) b. and I mean she was just impeccable (16R) c. and it just poured (16L) d. oh it's just out of this world (12R) Figure 6: Social class dierences in the use of just in four senses by Glasgow adults (% of group usage) MACAULAY 408 # Blackwell Publishers Ltd. 2002 There is nothing remarkable about this emphatic use but it does not occur in the working-class sessions. The middle-class speakers also use just with hedges, as in 6: 6. (Hedges in bold) a. Truro's really just a sort of market town (10L) b. to stop and just sort of pitch their camp there for the day (16R) c. it's really just a sort of buer (10R) d. it kind of had been programmed to really sort of just keep you in order (12R) There are not many examples of these uses in the middle-class conversations but there is none in the working-class conversations. In the adolescent conversations there are only minor social class dierences, with the middle-class adolescents using just with a frequency of 8.88 per thousand words and the working-class adolescents with a frequency of 8.30 per thousand words. The pattern and frequencies are remarkably similar to those in the adult conversations, showing that unlike very and quite, the adolescents are using the word in much the same way as the adults. There are even examples of the emphatic evaluative use (examples 7a and 7b) and of the use of just with hedges (examples 7c and 7d) that occur only in the conversations between middle-class adolescents: 7. (Hedges in bold) a. she's just dead annoying (5L) b. and I mean that's just stupid (5R) c. she's just sort of standing in for Mister Weir (3L) d. and eventually they just sort of ran out (4R) As with the adults, there are few examples but they occur only in the middle- class conversations. DISCUSSION Since the Ayr interviews and the Glasgow conversations were recorded under very dierent circumstances, their similarity is striking and raises questions as to why this situation should arise. However, as Bernstein pointed out, the diculty is `the problem of inferring from micro counts of specic linguistic choices to macro characteristics of the speech as a whole' (1971: 13). The remainder of this paper will be devoted to examining possible explanations for the social class dierence in the use of inected and other adverbs. It may be helpful to take as a starting-point three examples of empirical investigation. Bernstein (1971[1962] ) included adverb use in a list of features that he identied as characteristic of what he was then calling a restricted code rather than a public language: The restriction on the use of adjectives, uncommon adjectives, uncommon adverbs, the relative simplicity of the verbal form and the low proportion of subordinations ADVERBS AND SOCIAL CLASS 409 # Blackwell Publishers Ltd. 2002 supports the thesis that the working-class subjects do not explicate intent verbally and inasmuch as this is so the speech is relatively non-individuated. (Bernstein 1971: 109) It is not immediatelyobvious what `explicate intent verbally' or `non-individuated' mean or what role adverbs might play in either. Bernstein (1971) went on to explain the dierent character of an elaborated code: Individuated speech presupposes a history of a particular role relationship if it is to be prepared and delivered appropriately. Inasmuch as dierence is part of the expectation, there is less reliance or dependency on the listener; or rather this dependency is reduced by the explication of meaning. (Bernstein 1971: 113) In other words, `uncommon adverbs' help to make utterances more explicit. Labov, on the other hand, includes adverbs such as really as signals of intensity: `Intensity' is dened here as the emotional expression of social orientation toward the linguistic proposition: the commitment of the self to the proposition. (Labov 1984: 4344) Powell (1992), in a historical examination of the development of interpersonal and metalinguistic senses of stance adverbs, observes that certain adverbs can `act preemptively to inform and to persuade a hearer of the nature and importance of the speaker's evaluation' (1992: 76). She regards as interper- sonal use `any use which the OED denes as emphatic, emphatic use being preeminently expressive in function' (1992: 83). What evidence is there in the transcripts to support the hypothesis that the middle-class speakers use adverbs in -ly: (1) to be more explicit; (2) to express intensity; or (3) to signal the speaker's evaluation? The Oxford English Dictionary gives as its denition for the word explicit in relation to knowledge: `[d]eveloped in detail; hence, clear, denite.' In the Glasgow middle-class conversations there are examples of derived adverbs that might come under this heading, as in the examples in 8: 8. a. they're slightly dierent but they're exactly the same colour (10R) b. and it's immediately at the roadside (16L) c. it just goes downhill slowly (16L) d. a wee bit ambiguous here and there but generally okay (11L) In the examples in 8 the speakers appear to be trying to make the point clearly. There are, however, similar examples in the working-class sessions, as shown in 9: 9. a. you would just go along until you get to roughly the rst street (15L) b. there's really nothing to see in it but it's really quiet (18L) c. two completely dierent people (13R) d. he's aboot he's nearly as tall as taller than John must be aboot six two six four or something (14R) There are not many examples in either set of conversations and if this is what Bernstein meant by explicitness, then it does not appear to explain the social MACAULAY 410 # Blackwell Publishers Ltd. 2002 class dierence in the frequency of derived adverbs. I will return to the signicance of details later. As regards intensity, the examples in 10 are taken from the middle-class interviews in Ayr (Macaulay 1991: 125): 10. a. I found it extraordinarily boring (IM) b. I got absolutely sick of doing nothing (IM) c. but this zombie of a mother completely apathetic (WG) d. a terribly crippled bent old woman (DN) These examples support Labov's view of these adverbs as expressing intensity. There are 25 clear examples in the Ayr middle-class interviews but only three in the working-class interviews. Similar examples can be found in the Glasgow middle-class conversations: 11. a. and she was apparently absolutely horrendous (10L) b. who's got absolutely no sense of golng etiquette (11L) c. whereas the lady describing it thought it was absolutely perfect (11L) d you're either running around going d absolutely scatty chasing your tail or (10R) However, there are similar examples in the working-class sessions, though fewer, and most of the examples come from one man (18L): 12. a. I was there it was oh it was absolutely brilliant (18L) b. oh I mean it's amazing it's absolutely fantastic (18L) c. it seemed to me to be a perfectly good place (18L) d. everything's all just draining doon like that you know just completely totally unwinding (13R) So, there is some support for the view that the use of derived adverbs to express intensity contributes to the dierence in frequency between the two social classes. What about Powell's notion that adverbs are used to express the speaker's evaluation? The most obvious examples are evidentials, as in examples 13a and 13b but more important are examples such as the hedge in 13c and the intensier in 13d where the personal attitude is clearly stated: 13. a. but funnily enough I gave out (10L) b. interestingly enough that one of these programmes on the telly (16R) c. I thought technically it was brilliant but but it was boring (12L) d. San Francisco's actually quite chilly . . . it's it's amazingly chilly (16L) The examples in 13 are from the middle-class sessions. There are a few in the working-class conversations as well: 14. a. funnily enough we got that one (13R) b. but hmm unfortunately my trade went (18L) c. urban decay I mean you can get right into that politically (18R) d. but as I say I normally drink whisky (13R) ADVERBS AND SOCIAL CLASS 411 # Blackwell Publishers Ltd. 2002 However, there are few examples in either set of conversations, so the explanation for the dierence in frequency is unlikely to lie here. However, some of the adverbs used by the middle-class Glasgow adults suggest an attitude of condence in making categorical judgments that is less apparent in the working-class conversations: amazingly, awfully, badly, drastically, enormously, overly, properly and terribly. Even the other adverbs can have this eect when combined with adjectives as in the examples in 2, the kind of eect Louw (1993) describes as `semantic prosody.' 15. L10: mmhm her mother had me in stitches one day when I bumped into them in town and I think I'd had a particularly bad day with Kim and she was telling me all about her Fiona and what she was like at Kim's age who and she was apparently absolutely horrendous In 15 the adverbs emphasize the categorical judgments expressed by the adjectives. As was shown earlier the Glasgow middle-class adults use adjectives with a frequency of 34.16 per thousand words and the working-class adults with a frequency of 24.74. However, the dierence is greater when evaluative adjectives (Hunston and Sinclair 2000) are compared. The middle-class adults use evaluative adjectives with a frequency of 12.88 per thousand words and the working-class adults with a frequency of 8.67. Moreover, there is a dierence in the kind of evaluative adjectives used. In the working- class conversations 52 percent of the adjectives are simple words of approval or disapproval (e.g. good, bad, nice); in the middle-class conversations only 36 percent of the evaluative adjectives are of this kind. The middle-class adults use adjectives such as horrendous, horrible, hellish, chauvinistic, unattractive, messy, impressive, interesting, tremendous, fantastic, substantial and impeccable, but none of these or similar adjectives is used by the working-class adults. Of the three possible explanations examined so far, the use of derived adverbs to showintensity seems to be the most plausible thoughit may be dicult to separate this from evaluation. Moreover, it is not only derived adverbs that show this dierence. The dierence in the use of very is hardly one that can be explained in terms of register or education. It would also appear to be unrelated to explicitness. It is obviously used to express intensity and emphasis so that it can be used in the expression of evaluation, as in 16 where its use reects the speaker's assessment: 16. (Middle-class man) 16L: there's a very steep descent to it from the road . . . it's a very gradual descent to the bay . . . and it is actually a very nice walk Biber and Finegan in their cluster analysis of styles of stance found that the cluster that corresponds to `involved, intense conversational style' (1989: 110) was characterized by `frequent use of emphatics, hedges, and other general MACAULAY 412 # Blackwell Publishers Ltd. 2002 evidential markers' (1989: 111). Since the London-Lund Corpus of Spoken English (Svartvik and Quirk 1980) consists mainly of middle-class speakers and is the basis for Biber and Finegan's ndings it is not surprising that this description would also t the middle-class conversations in Glasgow. The middle-class Glasgow adults in general use more hedges than the working- class adults. For example, the middle-class speakers use sort of with a frequency of 1.84 instances per thousand words. The frequency for the working-class speakers is only 0.54 but even this is misleading because one of the working- class women uses sort of with a frequency of 2.3 per thousand words; the rest of the working-class speakers use sort of with a frequency of only 0.23. There is essentially no dierence in the use of kind of/kinda (MC 0.49 vs. WC 0.45). The middle-class Glasgow adults are also more likely to use you know in hedges (see Macaulay 2002). The middle-class Glasgow adults sometimes use derived adverbs along with hedges, as in the examples in 17: 17. (Adverbs in italics, hedges in bold) a. Truro's really just a sort of market town (10L) b. although the carpet was much thinner it was sort of badly tted (10R) c. and they were trying to sort of actually extend the ladies' rights (11L) d it kind of had been programmed to really sort of just keep you in order (12L) Biber and Finegan suggest that the certainty and emphatic forms in their conversational sample: seem to reect a sense of heightened emphatic excitement about the interaction, while the hedges seem to reect a lack of concern with precise details, indicating that the focus is on involved interaction rather than precise semantic expression. (Biber and Finegan 1989: 110) Biber and Finegan were interested in dierent styles employed in dierent genres, including written materials as well as spoken, so their emphasis is not on variation within conversational styles and cannot be expected to draw distinctions of this kind. Nevertheless, their conclusions are consistent with the middle-class Glasgow conversations. The question then becomes: what is it that characterizes the working-class speakers? One clue may lie in the phrase `a lack of concern with precise details' (Biber and Finegan 1989: 110) with reference to hedges in the London-Lund Corpus materials. It was apparent in the Ayr interviews that the working-class speakers were concerned about details. The most extreme example of this was Andrew Sinclair (Macaulay 1985, 1991: 249254): 18. I mean as one of thirteen of a family eh and I'm one of the oldest ones well there were four two boys and two girls older than me Here Sinclair takes care to make it clear just where he came in the birth order. He also gave many details of work in the coal mines. The Glasgow working-class speakers also include many details: ADVERBS AND SOCIAL CLASS 413 # Blackwell Publishers Ltd. 2002 19. (Conversation #13, working-class women) R13: and eh that's what happened there everybody was aw watching their bottles going doon you know doon and doon and doon the next thing oor table it was like a half bottle of vodka and a half bottle of whisky and six cans of Pils and th there was near enough another carry-oot was getting ordered L13: do you know you know that's what I would have ha had with me I wouldn't have had the vodka I'd have had like that my Pils maybe R13: aye L13: but I thought `Oh to hell I'm going I'm going to drink vodka tonight for a change' R13: aye but see that last one? the Times were gieing a can of Pils oot free in the Coop at the time can you mind o that? L13: oh right R13: so everybody was aw on Pils everybody that came in aw had aw these Pils they must have all been buying the Times and g giving giving aw these Pi cans of Pils L13: you were get you were getting you R13: cause their tables were full of them Everybody you could guarantee there was aboot six at each table aw drinking Pils and aw these cans were up and a big black bag at the the bottom of the hall aw the cans were getting put into cause that's what I was on an aw and then as I say we ended up going on to Haddows and getting mair This is a narrative about a night's drinking but nothing much happens in the story. Yet the details are important: the vodka, the whisky, the cans of Pils (beer). The evaluation comes in the line `there was near enough another carry- oot was getting ordered.' This means that despite the amount of drink on the table they were thinking of getting more from the o-licence (liquor store), and in the end they did: `we ended up going on to Haddows and getting mair.' It was clearly a night of prodigious drinking and the way it is communicated is through the details. There are no summarizing adjectives or adverbs. In the Ayr interviews, the lower-class speakers made use of several syntactic constructions that have a highlighting or intensifying eect (Macaulay 1991: 118123). The ve dierent constructions: demonstrative focusing; clefting; noun phrase preposing; left dislocation; and right dislocation are illustrated in20: 20. (Ayr, lower-class speakers) a. Demonstrative focusing i. that's us going for another game (WL) ii. and that was you shut in the house for a week (EL) MACAULAY 414 # Blackwell Publishers Ltd. 2002 b. Clefting i. it's a queer man and wife that doesnae have an argument (EL) ii. it's them that's running it now (MR) c. Noun phrase preposing i. an auld auld man he was you ken (WR) ii. and one of them he had been out with once or twice (AS) d. Left dislocation i. Mr Patterson he was a gentleman (WL) ii. but my own family they've had a lot of leeway (EL) e. Right dislocation i. she was a very quiet woman my mother (WR) ii. in fact he oered me a job Mr Cunningham (WL) The lower-class speakers inAyr used these constructions witha frequency of 2.91 per thousand words, in comparison to a frequency of 0.58 in the middle-class interviews. A similar dierence was found in the Glasgow adult conversations, withthe working-class speakers using these constructions witha frequency of 2.4 per thousand words and the middle-class speakers 0.23. The actual gures are shown in Figure 7. It can be seen that with the exception of demonstrative focusing, the middle-class speakers in both Ayr and Glasgow (columns 1 and 3) make some use of these constructions but very slight in comparison with the working-class speakers. This is the reverse of the situation with adverbs. Figure 7: Social class dierences by Ayr and Glasgow adults in the use of highlighting constructions (left dislocation, right dislocation, NP-fronting, it-clefting, demonstrative focusing) (freq. = per 1,000 words) ADVERBS AND SOCIAL CLASS 415 # Blackwell Publishers Ltd. 2002 CONCLUSION The examination of the Ayr interviews and the Glasgow conversations reveals a dierence in discourse style between the two social classes. The middle-class speakers appear to adopt two complementary strategies. One is to use adverbs (and adjectives) to make emphatic statements, making quite clear their opinions and their attitudes. The other is to soften their statements with hedges of various kinds. Both of these strategies are consistent with Biber and Finegan's conclusions from the London-Lund Corpus about an `involved, intense con- versational style' (1989: 110). The working-class speakers, on the other hand, seem to avoid these strategies and instead depend upon an accumulation of details and several movement rules to focus attention on certain constituents. Neither of these styles ts the kind of distinction between an elaborated code and a restricted code that Bernstein sought to draw from his investigation. Indeed, it could be argued that the working-class speakers are more explicit than the middle-class speakers. In the earlier paper (Macaulay 1995: 5153), I argued that the working- class use of quoted dialogue allowed the hearer more freedom to interpret the situation than did the use of evaluative adverbs and adjectives, which impose the speaker's interpretation on the listener. In the same way, the greater emphasis on details in the Glasgow working-class conversations provides the hearer with the information necessary to understand the situation. In contrast, the Glasgow middle-class speakers seem anxious to make sure that there is no doubt about their attitude or opinion (`very user-friendly really' L12; `it's actually quite nice for swimming' L16) and to do so they often employ adverbs. Given the dierent ways in which the data for the two studies were collected, the results cannot simply be the eect of the methodology. The consistency of the social class dierences is remarkable since there is nothing that stands out in the choice of topics that might aect the use of adverbs (or adjectives). Nor can the patterns of use be the result of interviewer bias, since there were no interviewers in the Glasgow sessions. One way of validating results is by testing again and again, by the same or dierent methods, in similar or dierent settings, with similar or dierent samples (Campbell and Fiske 1959). Since quantitative studies of discourse variation are not yet common it would be unwise to place too much signicance on the results of two small-scale studies, but the fact that the social class dierences show up so strongly in two quite dierent kinds of sample is some validation for the claim that something fundamental in the speech style used by the two social class groups in western Scotland governs their use of adverbs. MACAULAY 416 # Blackwell Publishers Ltd. 2002 NOTES 1. In the years in which I have been presenting papers on this topic I have received comments from a wide variety of scholars, including anonymous reviewers and the editors of this journal. To all of them I express my gratitude and my apologies for perhaps not beneting as much as I should have done from their help. 2. All page references are to the collected versions in Bernstein 1971. 3. For technical reasons three sessions were recorded with working-class women; one speaker was recorded twice with dierent interlocutors. As a result the number of participants in each social class/age/gender category is not totally consistent but since the results are presented in terms of frequencies, the dierence in absolute numbers need not materially aect any conclusions. REFERENCES Bell, Allan. 1984. Language style as audience design. Language in Society 13: 145204. Bernstein, Basil. 1959. A public language: Some sociological implications of a linguistic form. British Journal of Sociology 10: 311326 (reprinted in Bernstein 1971: 4160). Bernstein, Basil. 1960. Language and social class. British Journal of Sociology 11: 271 276 (reprinted in Bernstein 1971: 6167). Bernstein, Basil. 1962. Social class, linguistic codes, and grammatical elements. Language and Speech 5: 3146 (reprinted in Bernstein 1971: 95117). Bernstein, Basil. 1971. Theoretical studies towards a sociology of language (Class, Codes and Control, Volume 1). London: Routledge and Kegan Paul. Biber, Douglas and Edward Finegan. 1989. Styles of stance in English: Lexical and grammatical marking of evidentiality and aect. Text 9: 93124. Campbell, Donald T. and Donald W. Fiske. 1959. Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin 56: 81105. Coates, Jennifer. 1986. Women, Men and Language. London: Longman. Foulkes, Paul and Gerry Docherty (eds.). 1999. Urban Voices: Variation and Change in British Accents. London: Arnold. Hunston, Susan and John Sinclair. 1990. A local grammar of evaluation. In Susan Hunston and Geo Thompson (eds.) Evaluation in Text: Authorial Stance and the Construction of Discourse. Oxford: Oxford University Press. 75101. Labov, William. 1984. Intensity. In Deborah Schirin (ed.) Meaning, Form, and Use in Context: Linguistic Applications. Washington, D.C.: Georgetown University Press. 4370. Lee, David. 1987. The semantics of just. Journal of Pragmatics 11: 377398. Louw, Bill. 1993. Irony in the text or insincerity in the writer? The diagnostic potential of semantic prosodies. In Mona Baker, Gill Francis and Elena Tognini-Bognelli (eds.) Text and Technology: In Honour of John Sinclair. Amsterdam: John Benjamins. 157176. Macaulay, Ronald K. S. 1985. The narrative skills of a Scottish coal miner. In Manfred Gorlach (ed.) Focus on: Scotland. Amsterdam: John Benjamins. 101124. Macaulay, Ronald K. S. 1991. Locating Dialect in Discourse: The Language of Honest Men and Bonnie Lasses in Ayr. New York: Oxford University Press. Macaulay, Ronald K. S. 1995. The adverbs of authority. English World-Wide 16: 3760 (reprinted in Macaulay 1997). ADVERBS AND SOCIAL CLASS 417 # Blackwell Publishers Ltd. 2002 Macaulay, Ronald K. S. 1997. Standards and Variation in Urban Speech: Examples from Lowland Scots. Amsterdam: John Benjamins. Macaulay, Ronald K. S. 2000. Women talk about people, men talk about things. Paper given at IGALA 1, Stanford University, May 5, 2000. Macaulay, Ronald K. S. 2002. You know, it depends. To appear in Journal of Pragmatics. Macaulay, Ronald K. S. and Gavin D. Trevelyan. 1973. Language, Education and Employment in Glasgow. (Final report to the Social Science Research Council.) Edinburgh: The Scottish Council for Research in Education. Milroy, Lesley. 1980. Language and Social Networks. Oxford: Blackwell. Opdahl, Lise. 2000. LY or Zero Sux?: A Study in Variation of Dual-form Adverbs in Present- day English (two volumes). Frankfurt: Peter Lang. Powell, Mava Jo. 1992. The systematic development of correlated interpersonal and metalinguistic uses in stance adverbs. Cognitive Linguistics 3: 75110. Quirk, Randolph, Sidney Greenbaum, Georey Leech and Jan Svartvik. 1985. A Comprehensive Grammar of the English Language. London: Longman. Silverstein, Michael. 1996. Monoglot `standard' in America: Standardization and metaphors of linguistic hegemony. In Donald Brenneis and Ronald K. S. Macaulay (eds.) The Matrix of Language. Boulder, Colorado: Westview Press. 284306. Stuart-Smith, Jane. 1999. Glasgow. In Paul Foulkes and Gerry Docherty (eds.) Urban Voices: Variation and Change in British Accents. London: Arnold. 203222. Svartvik, Jan and Randolph Quirk (eds.). 1980. A Corpus of English Conversation. Lund: C. W. K. Gleerup. Tagliamonte, Sali and Rika Ito. 2001. Think really dierent: Continuity and specialization in the English adverb. Unpublished manuscript. Trudgill, Peter. 1974. The Social Dierentiation of English in Norwich. Cambridge: Cam- bridge University Press. Address correspondence to: Ronald Macaulay Pitzer College Claremont California 917116101 U.S.A. ronald_macaulay@pitzer.edu.