1、Chapter 7：Variation Studies An Early Study:Fischer(1958):(ng):,n p159 The choice between the ing and the in variants appears to be related to sex,class,personality(aggressive/cooperative),and mood(tense/relaxed)of the speaker,to the formality of the conversation and to the specific verb spoken.New Y
2、ork City:Labov(1966):incidence of r use pp161-165 Members of the highest and lowest social groups tend not to change their pronunciation after it becomes fixed in adolescence but members of middle social groups sometimes do because of their social aspirations.The r pronounced associated with the upp
3、er middle class even though members of that class do not always use such pronunciations,nor do they use them on all occasions.,Subjective reaction tests pp162-163 New Yorkers in the upper middle class and under the age of 40 almost unanimously approved r-pronunciation even though fewer than half act
4、ually used r in all possible instances.People below the age of 20 also used more r-pronunciation than people between the ages of 20 and 40,a fact that would suggest r-pronunciation to be on the increase.Above the age of 40,approval fell off to about 60 percent and use showed a dramatic decline to le
5、ss than 10 percent.The lower middle class speakers not only approve of r-pronunciation,they also tended to exceed what appear to be the norms for its use in the next highest class in reading word lists and in pronouncing minimal pairs of words.The amount of r use increases by social class and by for
6、mality of style.The lower middle class speakers out-perform his upper middle class speakers on word lists and pairs:an instance of hypercorrection.Lower middle class speakers know how prestigious r-pronunciation are and,when they are placed in situations which require them to monitor their speech cl
7、osely,they out-perform their reference group,the upper middle class.,Norwich and ReadingTrugill(1974)p167 Trugills analysis of the variable(ng),(t)and(h)shows that the higher the social class the more frequent is the use of,t and h variants in words like singing,butter and hammer rather than the cor
8、responding n,?and variants.Its use is related not only to social class but also to sex,with females showing a greater preference for than males,regardless of social class membership.Cheshire(1978)p168 Cheshire,(s)3rd person singular verb marking,constraints on usage:a verb stem always stakes the s w
9、hen it is used in the third person singular,the s ending is favored in all persons when the verb is a vernacular verb(commonly used verbs),but the s is not used at all if the verb has a complement in which the verb in the complement is marked for tense.High frequencies of s usage went with high inde
10、x scores and low frequencies with low index scores.(index based on ambition,degree of toughness and peer group status in order to assess the strength of an individuals membership in the boys vernacular culture.),Cheshire concluded that variation is controlled by both social and linguistic factors.In
11、 boys speech,variation is governed by norms that are central to the vernacular culture,and are transmitted through the peer group.Variation in the girls speech appears to be a more personal process and less rigidly controlled by vernacular norms.Both boys and girls are subject to two linguistic cons
12、traints on the form of regular present-tense verbs,of which one favors the use of the non-standard verb form and the other favors the use of the standard form.A further observation was that variation in the form of have and do appears to be due to linguistic changes in progress.”A topic will be expl
13、ained in detail in Chapter 8.,Detroit Study pp170-176 Shuy,Wolfram and Riley(1968):Whereas upper middle class speakers used multiple negation on about 2 percent of all possible occasions,the corresponding percentages for the other three social classes were as follows:lower middle class,11 percent,up
14、per working class,38 percent,and lower working class,70 percent.Although individuals exhibit a certain amount of inconsistency in their linguistic behavior,there is nevertheless a pattern to that behavior.Wolframs study(1969)attempted to show how the distribution of linguistic variables correlated w
15、ith such factors as social class,sex,age,and racial origin.Whereas it is quite possible that the differences between the two groups at each of the ends,i.e.,between the upper middle and the lower middle classes and between the upper working and the lower working classes,may not be significant,there
16、being only 12 subjects in each group,the difference between the top two groups as a whole and the bottom two groups as a whole,i.e.,between the middle class and the working class,almost certainly is and probably at a very high level of significance.Social status was the single most important variabl
17、e correlating with linguistic differences,with the clearest boundary being between the lower middle and upper working classes.,Macaulays study in Glasgow(1977)p173 A clear correlation between variation and social class.His two lowest classes to be much alike in behavior.With males,the greatest diffe
18、rence between classes was between his top class(professional and managerial)and the second highest class(white collar),whereas with females the greatest difference was between the two intermediate classes(white collar and skilled manual).Increase in age also seemed to be associated with an increase
19、in the difference between social classes,this difference showing itself to be clearly established in the 15 year olds surveyed(but apparently also in the 10 year olds).When individual rather group behavior was plotted for each variable,a continuum of behavior was exhibited in each case(the behavior
20、of certain individuals in each class overlapping the behavior of individuals in neighboring classes).We can conclude from his study that the linguistic behavior of individuals forms a continuum in the same way that social organization is continuous.Social classes are constructs imposed on this conti
21、nuum.If linguistic variation is correlated with the average behavior of individuals in these classes,it will show class difference.,Sankoff and Cedergren(1971)p174 The distribution of the variants of the(l)variable in Ontreal French is relate to both phonological and grammatical factors,not just soc
22、ial ones.The(l)is affected by its relationship to the following phonological segment and whether it occurs in either a personal or impersonal pronoun,when these are even of identical form,i.e.,il.Sankoff and Vincent(1977)pp174-175 Ne appears in contexts where speakers are most likely to be aware of
23、itself,and to be monitoring their own speech.The topic of language,instruction,discipline and religion tend to spirit people back to a normative world in which proper language becomes very salient.,Hudson(1996:178-180)pp175-176 There is no overlap within a particular sex,so that all members of the u
24、niversity-educated male group use less assimilation than all members of the next group,those with secondary education,and those,in turn,less than the men with primary education,and so on.There are overlaps between sexes,but even here the pattern is entirely consistent in that males always overlap th
25、e next lowest group of females.That is,if vowel assimilation is dispreferred,being associated with low educational attainment,males show the consequences of this just a little less than do females.,Belfast:Milroy and Milroy(1978),Milroy(1980,1987a)pp177-179 Social networks,which originate in kinship
26、 ties,determine an individuals access to employment and to other resources.People develop close and continuing relationships with each other,and they help one another,first their kin and then their co-religionists.How a stable set of linguistic norms emerges and maintains itself in a community.Lesle
27、y Milroy calls these vernacular norms,norms which are perceived as symbolizing values of solidarity and reciprocity rather status,and are not publicly codified or recognized.Each informant is placed on a six-point scale which characterized that persons participation in networks.The scale for scoring
28、 individual network strength used the following factors:membership in a high-density,territorially based cluster,kinship in the intermediate neighborhood,working with at least two people of the same sex from the same area,and voluntary leisure-time association with workmates.The Milroys examined 8 l
29、inguistic variables and found significant correlations between network strength and linguistic usage on five of these,two at p0.01(i.e.,there is less than one chance in a hundred that there is no such relationship)and three at p0.05(i.e.,here is less than one chance in twenty that there is no such r
30、elationship).,A closer inspection of the results by community showed that,with one exception,it was only in Ballymacarrett that there was a significant correlation between the variables and network strength.The greater the network strength the greater the incidence of the variants identified with th
31、e Belfast vernacular.A significant difference also exists between men and women in their use of the vernacular,with men showing a much greater incidence of vernacular usage.The stronger the social network,the greater the use of certain linguistic features of the vernacular.The results support Milroy
32、s hypothesis that a close-knit network has the capacity to function as a norm enforcement mechanism;there is no reason to suppose that linguistic norms are exempted from this process.Moreover,a close-knit network structure appears to be very common in low status communities.She adds that the close-k
33、nit network may be seen as an important social mechanism of vernacular maintenance,capable of operating effectively in opposition to a publicly endorsed and status-oriented set of legitimized linguistic norms.Once again,we see how low-status varieties of a language maintain themselves in the face of heavy competition from above:they enable those who use them to show their solidarity with one anoth
