Variation in the Use of ain't in an Urban British English Dialect Author(s): Jenny Cheshire Reviewed work(s): Source: Language

in Society, Vol. 10, No. 3 (Dec., 1981), pp. 365-381 Published by: Cambridge University Press Stable URL: http://www.jstor.org/stable/4167261 . Accessed: 18/03/2012 03:16
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access to Language in Society.

http://www.jstor.org

Lang. Soc. 10, 365-381.

Printedin the United States of America

Variation in the use of ain't in an urban British English dialect

JENNY CHESHIRE School of Modern Languages
Universitv of Bath

ABSTRACT

Ain't occurs as a sociolinguistic variablein workingclass speech in the town of Reading, England.The phonetic realizationsof ain't in Reading English etymologies, dependingmore on the syntactic do not accord with traditional environmentin which ain't occurs than on the standardEnglish forms from which they are usually assumed to derive. The phonetic variants are also markedfor semantic function in tag questions. Variationin the use of ain't can be explained as reflecting an ongoing linguistic change (sociolinguistics, dialectology, language change, theoreticallinguistics).
INTRODUCTION

variableain't, This paperanalyzes variationin the occurrenceof the grammatical using empiricaldata from a long-termstudy of the speech of three workingclass adolescentpeer groups in the town of Reading, England.' It describesthe syntactic and semantic functions that ain't fulfills in the speech of the peer groups and shows how these may be linked to the vernacularsubcultureto which the groups belong. It also considers the etymology of ain't and the linguistic changes that the feature has undergoneand that may still be in progress.
AIN'T IN NONSTANDARD READING ENGLISH

Ain't is a widespread feature of nonstandardEnglish dialects, both in Great form correspondsto several Britain and in the United States. The nonstandard standardEnglish verb forms. It occurs as the negative present tense contracted form of be, both as the copula:
I. We've got a park near us, but there ain't nothing over there.2

and as the auxiliary: 2. How come that ain't working? It is also used as the negative presenttense contractedform of the auxiliaryhave: 3. I ain't got one single flea in my hair, they're all married.
0047-4045/81/030365-17 $2.50 (?) 1981 Cambridge University Press

365

JENNY

CHESHIRE

It does not occur as the full verb have (this is usually replaced in colloquial speech by have + got). In some AmericanBlack English dialects it is also used as a past tense auxiliary, where standardEnglish has didn't: 4. I ain't see the fight, and I ain't hear the fight (from Labov et al. 1968). Whereasthe standardEnglish negative contractedforms are markedfor person
and number, aint't is not. Thus the one nonstandard form has five standard English equivalents: havetn't, hastn't, (a)mn not, aren't, and istI't.

DERIVATION

OF AIN

It is not unusual for the paradigmsin nonstandardEnglish dialects to contain fewer forms than the correspondingstandardEnglish paradigms.For example, the present tense verb system in standardEnglish consists of two forms: a suffixed form, used with third person singular subjects (e.g., he goes), and a suffixless form, used throughoutthe rest of the paradigm(e.g., I go, yvougo, we dialects in Britainand in the United States typically go, theYgo). Nonstandard have a simpler present tense system. In the English spoken in southwestern Britain, the suffixed form is used throughoutthe paradigm(Hughes & Trudgill 1979). In AppalachianEnglish the suffixed form is used with thirdperson plural subjectsas well as with thirdpersonsingularsubjects(Wolfram& Fasold 1974). Otherdialects, notablythe Black English Vernacular,and the English spoken in East Anglia, Britain, use the suffixless form throughoutthe paradigm.Thus the fact that the single form ain't occurs for standardEnglish (a)ln not, aren't, and isn't, on the one hand, and for standard English hasn't and haven't, on the other hand, can be seen as typifying a generaltrendin nonstandard dialects towardthe simplificationof linguistic systems. The fact thatthe one form, ain't, is used for two verbs that are quite distinct in meaningand in syntactic function, though, is an unusual phenomenonand one that requiressome explanation. There has been some discussion in the literatureconcerningthe derivationof ain't. In the past this discussion has centered on whether the form derives originally from the standardEnglish contracted forms of be, or the standard English contracted forms of have (see, for example, Willard 1936; Stevens 1954). In fact, however, the evidence thatis availablesuggests that ain't derived by regularsound change from the negative presenttense contractedformsof both have and be, so that the fact that the one form is used for both verbs is the result of a diachroniccoincidence.
Both Jespersen (1940) and Stevens (1954) describe the route by which ainl't

could have derivedfrom hasn't and haven't. In the Middle Englishperiod many of the irregularverb forms had two pronunciations,of which one predominated in positive sentences and the other in negative sentences. This meant that the contractednegative forms were clearly distinguishedfrom the positive forms, and there could be no confusion of meaning(as in the case, for example, of can 366

VARIATION

IN THE

USE

OF AIN T

[kenJ and can't [ka:ntl in British English today). The verb have, then, would have had two forms:one with a short vowel, used in positive sentences, and one with a long vowel, used in negative sentences. By the eighteenth century the fricative had been dropped in some contracted forms (in isn't, wasn't and doesn't, for example, as well as in hasn't and haven't), and also in poetry (cf. e'er, se'en). This is attestedto in the literatureof the time: Swift, for example, objected to the use of the form han't as early as 1710 (in The Tatler, no. 230). The initial aspirate would have been lost quite regularly in unstressed occurrences, and the long vowel would have diphthongized,via 1aeJI-[C]-e] to [efl), as it did, for example, in the compound verb form behave (Brook I958). Ain't, then, can be shown to derive regularlyfrom haven't and hasn't. However, the derivationof the form from the negative presenttense contractedforms thoughthere is evidence to suggest thatit derives of be is not so straightforward, from at least one of the contractedforms.
The most probable ancestor of ain't is the first person singular form am not.

The first nasal would have assimilated to the second, giving the form ann't (Jespersen 1940), possibly via an intermediarysyncopatedform amn't (cf. Stevens 1954), as has happenedin some Scottish English dialects. Simplificationof the long consonant may then have been accompanied by a lengthening of the vowel to Early Modern [ae, which diphthongizedto [ei , thus yielding the form
[eint.3

Jespersen suggests that ain't could also have derived independently from aren't and isn't. Loss of the [r] from aren't and subsequentdiphthongizationof the vowel would result in the form leinti. Loss of the fricative in isn't and a lowering and lengthening of the "unstable" vowel (as in don't and can't), followed by its subsequent diphthongizationwould also give the form [eintJ. However, there is little evidence to substantiatethese last suggestions, and the more likely explanation is that extension of the form leintl throughout the paradigmtook place by analogy with the other negative presenttense contracted forms (e.g., can't and won't), which have a single form throughout the paradigm. In some dialects the form [intl occurs as a phonetic variantof leInt]. It would seem reasonableto assume that the [intl variantis a survival of an earlier form and to expect it, therefore, to occur with thirdperson singularforms of the verb be (i.e., to correspondto standardEnglish isn't). In Reading English, however, the form lint] has a specific syntactic function. This will become clear in the analysis that follows.
AIN'T

AS A SOCIOLINGUISTIC

VARIABLE

IN READING

ENGLISH

The analysis of ain't in working class Reading English is based on the occurrence of the form in the informal, vernacularspeech of three adolescent peer groups. Membersof the peer groups were recordedover an eight-monthperiod in
367

JENNY

CHESHIRE

naturallyoccurringgroups in the adventureplaygroundsthat they used as meeting places. Care was taken to ensure that the recordingswere natural,spontaneous, and relaxed (see Cheshire forthcoming). In addition, recordingsof a more formalspeech style were obtainedfor some speakersby askingteachersto record membersof the peer groups at school, in the presence of the teacher. Thirteen boys and twelve girls were recordedat the playgrounds; they rangedin age from 9 to 17, thoughmost were aged between I I and 15. A totalof 331 occurrencesof ain't in the speech of the groups was analyzed, out of a potential439 occurrences of the form. Table I shows the frequency of occurrenceof nonstandardainl't in the verform occurs nacular speech of the three groups of speakers. The nonstandard with both the with a high frequency, and its distributionshows regularpatterning linguistic and the nonlinguisticcontext: all groups use ain't most frequentlyas the auxiliaryhave, and least frequentlyas the auxiliarybe, and the girls use ain't less often than the boys, in all cases. Ain't does not, however, show regularpatterningwith stylistic context. Table 2 shows the frequency indices for ain't as the negative auxiliary have and for ain't as the negative copula, in the vernacularstyle and the "school" style of seven of the boys. (The frequencyof occurrenceof the negative auxiliary be in ..school" style was too low to be includedhere.) Table 2 gives a clear indication of the absence of style shifting;in both cases the use of ain't actuallyincreasesin the more formal style.
TABLE I . Frequencyv indfices for ain't

(lift' t1ai

't=

0n111t

Group OrtsRoadboys Shinfield boys Shinfield girls

aux. have + not 91.18 100.00 65.58

full verb be + not 84.16 94.74 61.18

aux. be + not 79.07 63.16 42.11

TABLE 2. Frequency! indlices for ain't in vernacular st.vle ani(din 'school'' stv/e

Group frequency index: vernacularstyle aii't = neg.

aux .

Group frequency index: ''school'" style 100.00

77.78

93.02
74.47

ain't

te ho; neg.
copula

368

VARIATION

IN THE

USE

OF AIN'T

Linguistic constraints on ain't

Wolfram(1973) found thatthe occurrenceof ain't in the nonstandard English of Puerto Ricans in NewvYork City was constrained by the effect of multiple negation within the sentence and by the standardEnglish form to which ain't correspondedin the sentence. Ain't occurredmore often when it correspondedto standard English are + not and is + not thanwhen it correspondedto am + not. Neither of these constraints,however, has a significant effect on the use of ain't in vernacularspeech in Reading. Of far more significance in Reading English is the syntactic context in which ain't occurs. In the recordings, aiin't occurs in declarativesentences, in interrogativesentences, and in tag questions: 5. I ain't in the bloody guest house now. 6. Oh, ain't you going? 7. I'm going out with my bird now, ain't I? Interrogativesentences occurred only eleven times in the data, and therefore could not be included in the analysis. Table 3 shows the frequency of occurrence of ain't in declarative sentences and in tag questions. Bracketedfigures indicate that the numberof occurrences from which the index was calculatedis low, so that in these instancesthe indices are not statisticallysignificant. Despite gaps caused by low numbers, it is clear from Table 3 that the occurrenceof ain't is consistently higher in tag questions evident in the speech of the thanit is in declarativesentences. This is particularly girls. Tag questions, then, act as a strong constrainthere, favoring the use of the form. nonstandard

TABLE 3. FrequencY of occurrence of ain't in dec-larative sentences and in taig questions

Declarative sentences
Aux. have

Tag questions 100.00 (100.00) 80.00 95.00 100.00 100.00 89.66 96.30 100.00 (75.00) (66.67) 88.89

Orts Road Shinfield boys Shinfield girls Total Main verb be Orts Road Shinfield boys Shinfield girls Total Aux. be Orts Road Shinfield boys Shinfield girls Total

88.68 100.00 62.50 78.57 68.00 92.31 44.44 59.83 76.67 60.00 3 1.25 60.66

369

JENNY

CHESHIRE

Phonetic realizations of ain't

The discussion so far has used the general term ain't to refer to all the nonstandardrealizationsof the negative presenttense contractedforms of the copula and of the auxiliaries be and have. In reality, however, the nonstandard form has a range of phonetic realizations. The Survey of English Dialects shows ain't as being variouslyrealizedas leintl and lentl in Berkshire(Orton,et al. 1968). The informantsused in this survey, however, were all from ruralareas, and there are no observationsof the English spoken in urbancenters in Berkshire. In the data obtainedfor the presentstudy the realizationsof ain't fall into two groups:those thatcan be said to approachain't (including leint], [ein], [ei?], [Lil, [ei?J, lent] and Lenl), and those that can be said to approachin't (including lint], [in], and LI]).In addition, there were three occurrencesin the data of the form lant], and three of the form [;nJ, all in tag questions. Two of these occur as the negative auxiliary be, and four as the negative copula, and all persons of the verb are involved, with the exception of the first person singular. However, these lantl forms are too low in numberto allow furtheranalysis. If we take the traditionalview that ain't derives from the standardEnglish forms, then it would be reasonableto expect thata phoneticrealizationapproaching in't would occur with third person singularsubjects of the copula, or of the auxiliarybe - in otherwords, thatnonstandard would correspondto standard itm't English isn't, and nonstandardain't to standard English (a)m ttot, arenl't,
hasn't, and haven't.

Table 4 shows the distributionof itm't (i.e., of those phonetic realizationsthat contain the vowel III), with third person singular subjects and with non-third person singularsubjects, for all verbs. The indices for the Shinfield boys' group ini'tcorrespondsto show the distributionthat would be expected if nonstandard standardisn't: ini't does not occur as the negative auxiliary have, nor does it

TABLE 4. FrequenmY ind(1ex of in't

fr no. of in'! f s x

no. of in't + no. of aini't forms Auxiliary have 3s Orts Road Shinfield boys Shintield girls Total 24.((X) (0.0() (.(X) 16.22 other subjects 8.33 (M.00) 0.00 5.17 3s 63.24 33.33 61.54 60.34 Main verb be other subjects 41.67 0.00 10.00 19.35 3s 50.00 (66.67) (50.00) 53.85 Auxiliary be other subjects 15.38 0.00 (25())) 12.82

370

VARIATION

IN THE

USE

OF AIN'T

occur with non-third person singular subjects of the negative copula and of the negative auxiliary be. However, this group of speakers is small, consisting of only threeboys. The other, larger,groups do not confine theiruse of ain't to this environment. The Shinfield girls use in't for the negative copula and for the negative auxiliarybe with subjectsthatare not thirdperson singular,and the Orts Road boys use in't not only with non-thirdpersonsingularsubjectsof the copula and auxiliary be but also as the auxiliary have, with all subjects. The most favorable environmentfor the occurrenceof in't for all groups of speakers is with third person singular forms of the verb to be, as we would expect. However, in Reading English the use of in't is not confined to this environment, as we have seen, and the nonstandard forms ain't and in't, as a result, do not correspondexactly to their supposed standardEnglish equivalents. This becomes clearer when we consider the phonetic realizationsof ain't in the syntactic contexts in which it occurs most often. Table 5 and Table 6 show the frequencyindices for the main phonetic realizations of the nonstandardl form in declarativesentences and in tag questions. Table 5 shows the frequency indices with third person singular subjects, and Table 6 shows the frequency indices with non-third person singular subjects. In both tables the figures representthe total use of each form by all speakersin the study. (A full version of Table 5 and Table 6, showing the breakdownfor each of the threepeer groups, can be seen in the Appendix.The groupfigures show the same distributionof forms, except that the Shinfield girls use the standardEnglish contractedform more often than ain't or in't, with third person singular subjects.) Table 5 shows that although in't forms do occur more often with thirdperson singular subjects and the verb be, as we would expect if in't corresponds to standardEnglish isn't, these occurrencesof in't are almost entirely in tag questions. In declarative sentences the form ain't predominates,for all verbs, and in't occurs only rarely (as the copula and the auxiliary be, and never as the auxiliary have). In tag questions, the distributionis reversed: in't predominates

TABLE

5.

Percentage use of ain't, in't, and standardcforms, with third person singular subjetts

Declarative sentences 3rd pers. sing. subjects aux. have main verb be aux. be ain't 82.35
ain't

Tag questions ain't 33.33

ain't

ill t 0.00
in t

hasn 't

in't 66.67
in't

hasn't 0.00
isn't

17.65
istn t

54.32 50.0()

6.17 8.33

39.51 41.67

2.86 0.00

92.86 85.71

4.28 14.29

371

JENNY

CHESHIRE

TABLE

6.

Pe r entageuse of in 't, ain 't and standard forns. Declarative sentences

uith noni-third person sintgular subje(cts Tag questions

Non-3rd pers. sing. subjects aux. have niain verb he aux. be

aini't

ill't

havent't

aini't

inj't

haven't

76.56 ain't 58.33 61.22

0.00 in't 0.00 0.00

23.44
aretn't

41.67 38.78

72.73 ain't 40.00 44.44

27.27 In't 60.00 55.56

0(X)
areni't

(.00 0.00

here, for all verbs, and ain't occurs much less often as auxiliaryhave, and rarely or never as the auxiliary be or as the copula. Table 6 shows that these tendencies persist when the subject is non-third person singular. If the nonstandard forms correspondedto the standardEnglish forms, then in't would not be expected to occur at all in this table. Yet ini't is again the predominant form in tag questions when the verb is the auxiliary be or the copula, and it also occurs, though to a more limited extent, when the verb is the auxiliary have. In declarativesentences, ain't is the only nonstandard form that is used, for all verbs. It is clear, then, that the nonstandard forms imi't and ain't do not stand in a simple relationshipto the standardEnglish forms isn't and areni't. That they do bear some relationshipto the standardEnglish verb forms is shown by the fact that in't forms occur with the verb have less frequentlythanwith the auxiliarybe or with the copula. And in the environmentthat is, so to speak, the farthest removedfrom isn't (i.e., non-thirdsingularsubjectsand the verb have), alint't is the preferredform in tag questions, ratherthan in't, although inl't does, even here, occur 27.27% of the time. In all otherenvironments,the mainconstrainton the nonstandard variantsis the syntacticcontext in which they occur: ain't is the form in tag preferredform in declarativesentences, and itl't is the predominant questions. The tendencyto use in't as an invariantform in tag questions is not confinedto Scots speech ReadingEnglish. Brownand Millar(1978) reportthat in Edinburgh also occurs with all subjects of the verb be. An invariantpast tense form, itm't win't, also occurs, with all subjects, in some varieties of Scottish English.4
STANDARD ENGLISH FORMS

Although the standardEnglish forms are not of central interesthere, it is worth noting the forms that are preferredin vernacularspeech. In colloquial standard English, contractednegatives are abbreviatedin one of two ways. The verb may be assimilated to the preceding subject:
372

VkRIATION

IN THE

USE

OF AIN

8. You're not making room for me. or the particle may be reduced and assimilated to the preceding verb: 9. You aren't a virgin. In the data, the forrner,uncontractednegative form predominateswhen the verb is the auxiliary be or the copula. It occurs ioo% of the time with the auxiliary be and 74% of the time with the copula. With the exception of isn't, however, and one isolated occurrence of haveni't, the standardEnglish forms occur only in declarativesentences. Possibly the use of the full negative particle here emphasizes the negation of the verb. With auxiliary have, however, the general tendency of southern English dialects to prefer the contractednegative form is followed: lo. No, you've had one. I haven't even had one, so shut up. Wherethe standardEnglishforms are used in vernacularspeech, they are used in the same way as they are in the standardlanguage;that is, isn't occurs only with third person singular subjects of the verb be, haven't only with non-third person singularsubjects of the verb have, and so on.

THE

SEMANTIC

FUNCTIONS

OF TAG

QUESTIONS

It is instructiveat this point to look at the semanticpropertiesof the tag questions that are used by the speakers. The syntactic and semantic structureof tag questions has been discussed in some detail (e.g., Palmer 1965; Arbini 1969; Huddleston 1970; Quirk, Greenbaum, Leech, & Svartvik 1972; Cattell 1973; Millar & Brown 1979). For our purposesthe most relevantanalysis is that given by Hudson (1975). This will be described in some detail here in order to show how it applies to the data from the speech of the peer groups. Hudson analyzes polar questions (consideringtag questions as reduced polar interrogatives)in terms of their syntactic, semantic, and pragmaticproperties. The pragmatic analysis is based on the concept of "illocutionaryforce," as conceived by Austin (1962), and on Grice's rules of social interaction(Grice 1975). Hudson shows that the illocutionarymeaningof an utteranceis relatedto the syntactic properties of the sentence via an intermediaryset of semantic propertiesthat are concerned with the situation of the speaker and the listener, and with their beliefs regardingthe 'truth"of the propositionexpressed in the sentence. He points out, for example, that although sentences I1 and 12 below have differentsyntactic properties,they both can have the 'illocutionary force" (in the sense used by Austin 119621)of WARNING, in certain circumstances:
II.
I2.

That kind of lock isn't safe. (from Hudson 1975)

Is that kind of lock safe? (from Hudson 1975)

373

JENNY

CHESHIRE

Sentence l i will serve as a warningonly if the heareris in dangerbecause of the lock and not if said, for example, by A to B while they are looking at a catalog of locks. For sentence 12 to serve as a warning, not only must the hearer be in danger because of the lock, but he must also believe that the lock is for dangerous. It would be inappropriate A to use sentence 12 to B if B knew nothing about locks, and if A knew that B knew nothing, althoughsentence i in would be perfectly appropriate these circumstances. Conversely, sentence x3 can be used both as a challenge and as a kind of apology, depending partly on the intonation contours of the main clause and partlyon the circumstancesin which it is uttered. When spoken with a falling intonationon the main clause and a rising intonationon the tag, the sentence acts as a challenge. When spoken with a rising intonationon the main clause, the sentence acts as an apology. 13. This is your seat, is it'?(from Hudson 1975) In this case, the syntacticpropertiesof the sentence are identicalin bothcases (although the intonationis different), but the beliefs of the speakerconcerning the relationship of the hearer to himself (and of the hearer to the seat!) are different. force" of a sentence, Hudson draws a distinction between the "'illocutionary which may vary on the differentoccasions on which it is uttered, and the "permanent" propertiesof the sentence that are relevantto the syntactic distinctions of mood (e.g., declarative, interrogative,imperative)and that apply at the level of semantic structure. These "permanent properties" of sentences can be specified in termsof 'sincerity conditions." Declarativesentences, for example, are subject to the sincerity condition below: The speaker believes that the propositionis true. And interrogativesentences are subject to the sincerity condition: The speakerbelieves that the hearerknows at least as well as he himself does whether the propositionis true or false. Hudson's discussion of the meaning of questions, then, involves the following separatecategories:syntacticcategories(e.g., interrogative,declarative),semantic categories (e.g., question, statement),and sincerityconditionson the semantic categories. The total meaningof the question, of course, also involves general pragmaticprinciples(e.g., Griceanrules of conversation)and illocutionaryforce '("whateverconclusions the hearer may draw from a particularutteranceon a occasion" 1301). particular Sentences that contain tag questions are subjectboth to the sinceritycondition sentences, on declarativesentences and to the sincerityconditionon interrogative since they contain both a statement(in the main sentence) and a question (in the tag). The majorityof the tag questions used by the peer groups in the data do
374

VARIATION

IN THE

USE

OF AIN T

conform to both these sincerityconditions. Tag questions are most often used to seek confirmationof an offered fact:
14. He lives here, doesn't he?

of to seek corroboration a statement:

I5.

My brothercarried him all the way down the hospital, didn't he?

and to seek supportfor an opinion: i6. Well, it is rude, isn't it'? In all of these examples, the tag is spoken with a falling intonation.The function of these tag questions corresponds to what is generally considered to be the central function of tags: to seek confirmationor corroborationfor the hopes or suppositionsexpressed in the sentence to which they are attached(cf. Stockwell,
Schachter, & Partee 1973).

As noted above, the sincerity conditions on both declarative sentences and interrogativesentences are fulfilled in these tags; the speaker believes that the statementin the main sentence is true, and he believes that the hearerknows "at least as well as he does" whether it is true or false. In addition, he expects the hearerto confirm the propositionof the main sentence. There are a numberof tag questions in the data, however, that do not expect the hearer to confirm the proposition of the main sentence, and that do not conform to the sincerity conditions on interrogativesentences. The interchange below provides one example: Mick: Any intruderthat comes down here, they gets beat. Jennv: Why didn't you beat me, then? Mick: 'Cos you're a girl, in't you? Here the propositionin Mick's main sentence, you are a girl, is obviously true, and the function of the tag is not to ask for confirmation,but ratherto show that the question Mick had been asked was in his opinion a foolish one. The tag carries overtones of sarcasmand of slight hostility; no answer is requiredhere. The sinceritycondition on interrogativesentences does not apply; the speakeris fully committedto the truthof the proposition, and he assumes, in addition, that the heareralso knows that the propositionis true. A furtherexample is Cathy's tag in the conversationalfragmentbelow: Jackv: We're goitng to Southsea on the seventeenth of next month. And on Sunday they... Cathv: Yeah, and I can't bloody go. Jenny: Why not? Cathy: 'Cos I'm going on fucking holiday, in I? Again, no answer is expected, and it would in any case have been impossible to provideone, since I had no way of knowing when Cathy was going on holiday,
375

JENNY

CHESHIRE

and she knew that I did not know. The sincerity condition on interrogative sentences cannot apply here. The speaker believes that the proposition in the main sentence is true, but she does not believe thatthe hearerknows whetherit is true or false. Tag questionsof this kindoccur regularlyin workingclass speech. They occur in Reading, as we have seen, and they have been noted in Edinburgh(Millar& Brown 1979) and in London (Hudson 1975). Hudsonsuggests that it is because the sincerity conditions are not fulfilled that tags of this type strike outsidersas odd. Millar and Brown suggest furtherthat the aggressive and hostile overtones that these tags possess result from the fact that they "'playon the conventional meaning" associated with tag questions that have the same syntactic structure. of They "mimic the presentation an analytictruth,such thatthe heareris madeto feel that he really should have known either by intuition, perception,or deduction that the propositionwas true" (35). A furthertype of tag question, whose occurrenceis not confined to working class speech, occurs in the data, and again no answeris requiredfrom the hearer: 17. You're a fucking hard nut, in't you'? This tag question was used provocatively, with the intentionof startinga fight. Although no verbal answer was expected, a physical response may well have been; in this case, the hearerimmediatelyjumped on the speakerand threw him to the ground. It is difficult to establish whether the sincerity conditions are fulfilled here or not; the speakermay not really believe that his friend is a "hard nut," but may believe it only temporarily,or may be pretendingto believe it in order to start a fight, and his assumptionsabout the hearer's beliefs are equally unclear. Broadlyspeaking, then, the tag questionsthatare used by the speakersfall into two groups:a largergroup, where tags have the "regular" functionof requiring and where Hudson's sincerity conditions are fulconfirmationor corroboration filled, and a smaller group, consisting of tags that do not requirean answer and that do not fulfill Hudson's sincerity conditions.5 These groups turnout to be extremely relevantto our analysis of the phonetic
realizations of ain1't. Table 7 shows the number of times that ain't, im1'tand the

standardEnglish verb forms occur in the vernacularstyle of all speakers in the " .wo groups(or, types) of tags. Group I consists of those "conventional tags that requirean answerandthatfulfill the sincerityconditions. Group2 consists of what tags: tags that do not requirean answer and nay be termed "nonconventional'" thatdo not fulfill the sincerityconditions. Both groupsof tags are spoken with a falling intonation. Table 7 shows clearly that, whereas iti't occurs in both conventional and Englishcontractedverb formsoccur tags, aiin't and the standard nlonconventional only in conventional tags. The use of in't, in other words, is categorical in nonconventionaltags, but is variable in conventional tags.
376

VARIATION

IN

THE

USE

OF

AIN'T

TABLE 7. Nunmer oj occurrences

in "cotventitotial"'

of in't, ain't at1il standardl English f0rlnns tags

tags andi in ''nonconventional''

standard ill't Tag type Non-3s subjects copula 1 2 1 tainl't 2 l English form 2

aux. be aux. have

TOTAL 3s subjects copula

1
7

4 2
7

4 7
12

0 0
0

0 1
1

0 0
0

59

aux. be aux. have

TOTAL Combined total

3 6
68 75

3 0
8 15

0 3
5 17

0) 0
0 0

0
4 5

0 0
0 0

(all subjects)

There is an interestinglink here between the categoricaluse of inl't in nonconventionaltags and the vernacular culturein which the speakersparticipate.Nonhave seen, have certainsemanticpropertiesin common: conventionaltags, as wve they are nonconducive,andthey do not conformto Hudson'ssincerityconditions. In addition, they all convey overtones of aggression, assertion, or hostility to the hearer. Almost all the members of the peer groups were involved in a street-corner vernacular culture, which is in many respectsdirectlyopposed to the mainstream culture in society. Dominantthemes within the culture are aggression, violence, and hostility. These can be seen as underlyingfeaturesof the shoplifting, fighting, arson, and vandalism that the peer groups enjoy. Prestige within the peer group is achieved throughsuccess in these activities, by carryingweapons, and by generally "'actingtough.'" The extent to which speakersparticipatein the vernacularculture is reflected in their language by the frequency with which they use certain nonstandard linguistic features (Cheshire forthcoming). Interestingly, ain't is one feature whose frequencyof occurrenceis not directly correlatedwith the degree to which speakers adhere to the vernacularculture. It seems that there is a more indirect link between aiii't and the vernacularculture: when the feature occurs in a nonconventionaltag, carryingovertones of aggressionor hostility, then the form in't is categorical;when it occurs elsewhere, then any of the variableforms may occur. We can say, then, that when tag questions overtly specify the dominant themes of the vernacular culture, in't is used as an invariantnegative form of be
377

JENNY

CHESHIRE

and auxiliary have (other auxiliaries, such as can or will, may also occur in nonconventionaltags). This kind of phenomenon is not unknown in language. In his discussion of negative concord in the Black English Vernacular,Labov (1972) gives an example of a culturalfeature [- book learning]coinciding with the use of nonstandard grammar.And Hudson (1975) argues that the n't thatoccurs in exclamationsof this kind,
19. Hasn't he gone a long way'?

is not a reflex of the deep formative NEG, as it is in questions such as this:

20.

Hasn't he finished yet'?

(though it presumablywas at one stage of English), but is instead a markerof "'exclamation,'" relatedto the negative form nt' only at the level of morphology. Hudson points to other "syntactic splits" that have occurred in English - for example, the split between iuse and the modal auxiliaryused or the split between owe and ought (ought was at one time the past tense form of owe). It could be similarlyarguedthatthe iju'tthatoccurs as an invariantform in nonconventional tag questions is not related to the third person singular present tense form im1't other than historically and that it is now a markerof the vernacularthemes of
aggression and hostility (a "force marker"; Hudson 1975).6

ONGOING

l INGUISJ

IC CHANGE

IN THE

FORM OF AIN'

As we have seen, the phoneticvariantsof aini't are best explained as the resultof the combined effects of linguistic change and morphologicalanalogy. The data suggest thatthese processes are still at work on the phoneticrealizationsof ain't. It is reasonableto assume that the form in't derived by regularsound change from standardEnglish isn't. In the tag questions in the data, the verb be occurs more often than the verb hav'e (105 times, as comparedto 22 times), and third person singularsubjects occur more often than other subjects (82 times with be and io times with hai'e). This means that third person singular torms of be account for 78% of the verbs in tag questions. If this is typical of vernacular speech in Reading, and there is no reason to suppose that it is not, then this means that the form it1't occurs more often in tag questions than elsewhere. It seems highly plausible that, as a result of its widespread occurrence in tag (luestions, it is becoming used, by morphologicalanalogy, with subjects other thanthirdperson singularsubjects, and with auxiliaryhave as well as with be in tags. In other words, what may be happeninghere is precisely what is presumed to have happened at an earlier stage of English to the form ain't: a form that occurs frequently in one morphological environment in colloquial speech is spreadingthroughoutthe paradigm,though here it is happening,at present,only
378

VARIATION

IN THE

USE

OF AIN

of in tag questions. In addition,the form in't appearsto have become a marker an overt vemacular norm in nonconventionaltag questions, so that here it occurs . categorically We can use the figures for the frequency of occurrenceof ir7't in the data to guess at the state of progressof the change. Table 5 suggests that the change is furtheradvanced with third person singular forms of the verb have: in't occurs 66.7% of the time. Again, this could be due to the fact that thirdperson singular subjectsoccur more often thanother subjects. Table 6 indicatesthat the change is also spreadingto non-third person singularforms of be, and that it is beginning to affect non-third person singular forms of have also.

CONCLUSION

This analysis of the use of ain't in working class speech in Reading has shown that, whereas in standardEnglish the negative presenttense contractedforms of have and be are markedfor subject and for the verb, in nonstandardReading speech they are marked instead for syntactic function: ain't is used predominantly in declarativesentences, and in't is used predominantlyin tag questions. In nonconventionaltag questions the use of in't is categorical, and this may be due to the fact thathere it is a markerof an overt vemacularnorm [ +aggressionJ. In conventional tag questions the use of in't is variable, but there is some indicationthat a linguistic change is in progresshere, towardthe use of in't with all subjects and with both have and be.

NOTES

I am very grateful to Richard Hudson and Peter Trudgill for their comments on an earlier x. version of this paper. Unless otherwise stated, all sentences used as examples are from the speech of membersof the 2. peer groups. 3. Alternatively, in some British English, eastern New England, and coastal southernAmerican English dialects, Early Modern laei before a following [ntl may have developed a schwa glide, or lengthed to producethe form ia:nti. This would account for the hithertounexplainedoccurrenceof the form la:nt] with first person singular subjects in interrogativesentences in standardBritish English and also in declarative sentences in some nonstandarddialects. The orthographyaren't would be explained as dialect borrowing or as analogy with other members of the paradigm
(McDavid 1941).

A simplerexplanationwould see the occurrenceof [ct:nt]with first person singularsubjects as the result of analogical extension of other forms in the paradigm. This is confirmed by the researchof John Kirk of the University of Sheffield. 4. For a more rigorousdescriptionof the syntactic and semantic structureof tag questions, which 5. includes their functions in working class speech, see Millar & Brown (1979). It could be furtherarguedthat nonconventionaltags of the type described here are not tags in 6. either a semantic or a syntactic sense. They do not have the semantic propertiesof tag questions (in thatthey are not conducive and do not conformto Hudson'ssincerityconditions), nor do they have the of syntactic structure tag questions, which normallyrepeatthe NP and the VP of the main sentence.

379

JENNY

CHESHIRE

APPENDIX of Table 6 (realizatiotns ain't) Full verfsion Table 5 and(I of Declarativesentences

3rl person sinrgulair aim'J int'I hastn't aim't

Tag questions
inl'I hasti '

hav'e Orts Road

Shinfield boys

100.00
(100.00)

0.00
(0.00)

0.00
(0.00)

14.29
-

85.71
-

0.00
-

Shinfield girls TOTAL main verb be Orts Road Shinfield boys Shinfield girls TOTAL auxiliary be Orts Road Shinfield boys
Shinfield girls

60.00 82.35 62.50 83.33 40.00 54.32 80.00 (50.00)

20.00

0.00 0.00 7.50 0.00 5.71 6.17 0.00 (0.0())

20.00

40.00 17.65 ist,',t 30.00 16.67 54.29 39.51

is,i't

(100.00) 33.33 0.00 (25.00) 3.85 2.86 (0.00) (().00)

-

(0.00) 66.67 100.00 (75.00) 84.62 92.86 (10(.0(0) (66.67)

-

(0.00) 0.00 is,, I 0.00 (0.00) 11.54 4.29 ispi'I (0.00) (33.33)
-

20.00 (50.00)
60.00

TOTAL Non-3rdl perso.i sinigular have Orts Road

Shinfield boys

50.00

8.33

41.67

0.00

85.71

14.29

hav-en1't

82.86
(10(0.00)

0.00
(0.00)

17.14
(0.00)

57.14
(100.00)

42.86
(0.00)

c'I hve(: 0.00

(0.00)

Shinfield girls TOTAL main verb be Orts Road

Shinfield boys

64.00 76.56 60.00

0M.00

0.0() 0.00 0.00

0.00

36.00 23.44 aren't 40.00

0.00

(66.67) 72.73 16.67

(100.00)

(0.00) 27.27 83.33

(0.00)

(33.33) 0.00 aren't 0.00

(0.00)

Shinfield girls TOTAL

auxiliary be

42.11 58.33 76.00 61.54 27.27 61.22

0.00 0.00 0.00 0.00 0.00 0.00

57.89 41.67
aretn't

(50.00) 40.00 42.86 (100.00) (0.00) 44.44

(50.00) 60.(0) 57.14 (0.0()) (0.00) 55.56

0.00 0.()
airetn't

Orts Road Shinfield boys Shinfield girls TOTAL

24.00 38.46 72.73 38.78

0.00 (0.0() (0.0() 0.00

REFERENCES

Arbini, R. (1969). Tag-questions and tag-imperatives in English. Journal of Linguistic.s 5:205-14. Austin, J. L. (1962). How to do thitngs with wortls. Cambridge, Mass.: Harvard University Press. Brook, G. L. (1958). A historn of the Etnglishlanguage. London: Andre Deutsch. Brown, K., & Millar, B. (1978). Auxiliary verbs in Edinburghspeech. University of Edinburgh, Departmentof Linguistics:Work in Progress 11.
Cattell, R. (1973). Negative transportation and tag questions. Language 49(3):612-39. Cheshire,J. (forthcoming). Vari(ation in an English(lialect:A sociolinguisticstudY.CambridgeUni-

versity Press.

380

VARIATION

IN THE

USE

OF ANV T

Grice, M. P. (1975). Logic and conversation. In P. Cole & J. Morgan(eds.), Stax.r andi semantics 3: Speech acts. London: Academic Press. Huddleston, R. (1970). Two approachesto the analysis of tags. Joutrnatil Linguistics 6:215-21. of Hudson, R. A. (1975). The meaning of questions. Laut,t,gage 51(1): 1-31. Hughes, G. A., & Trudgill, P. J. ( 1979). Englislzaccents and dialet-is: An introductionto regional and so(ial varieties of British Englishl. London: EdwardArnold. on Jespersen, 0. (1940). A motlern English gramnmttar historical principles. Part V. Copenhagen: EjnarMunksgaard. Labov, W. (1972). Negative attractionand negative concord in English granimar. Latnguage 48: 773-8 18. Labov, W., Cohen, P., Robins, C., & Lewis, J. (1968). A situdv the nonistand(ar(d of Etnglish Negro of and Puerto Rican speakers in New YorkCity . I & i1. Final report,Co-operativeResearchProject 3288. Washington, D.C.: U.S. Office of Health, Educationand Welfare. McDavid, R. 1. (1941). Ain't I and aren't 1. Lainguiage 57-59. 17: Millar, M., & Brown, K. (1979). Tag questions in Edinburghspeech. Linguisti.scheBerichte 6o: 24-45. Orton, H., et al (I968). Survev of Eiglish Dialects. Vol. IV, part 3. Leeds: E. J. Arnold. Palmer, F. R. (1965). A linguistic study of the English veerb.London: Longman. Quirk, R., Greenbaum.S., Leech, G., & Svartvik, J. (1972). A grammnar Etglish. ofcontemnporary London: Longman. Stevens. M. (1954). The derivationof ain't. AmericanspeeCh 29: 196-201. Stockwell, R. P., Schachter, P.. & Partee. B. M. (1973). The major syntactic structuresof Etnglish. New York: Holt, Rinehart& Winston. Willard, E. P. (1936). The Origitnof ain't. Wordstiudv XI(2). Wolfram, W. (1973). Sociolinguistic aspects of assimilation: Puerto Rican,English imn New York City. Arlington, Va.: Center for Applied Linguistics. Wolfram, W., & Fasold, R. (1974). 7he stud! of social dialects it Amnerician Etnglish.Englewood Cliffs, N.J.: Prentice-Hall.

381