Peasant Men Can't Get Wives: Language Change and Sex Roles in a Bilingual Community Author(s): Susan Gal

Source: Language in Society, Vol. 7, No. 1 (Apr., 1978), pp. 1-16 Published by: Cambridge University Press Stable URL: . Accessed: 27/12/2010 13:39
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . . Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. 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

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

Lakoff I975).H. Hammel for their many suggestions. Anthropology Traineeship at the University of California. I . Farber I974).M. Rather than simply isolatinga linguistic correlateof sex. in choice of lexical items (Swacker I975). sexual differentiation of speech is expected to occur whenever a social division exists between the rolesof men and women . My thanks to Paul Kay. in choice of languageby bilinguals(Rubin 1970.recentwork has shown that linguistic differencesbetween men and women can appearat variouslevels of grammar:in phonology (Anshen I969. Sankoff& Cedergren I97I). social determinants of language shift.Further.Berkeley ABSTRACT Languageshift from German-Hungarianbilingualism to the exclusive use of German is occurring in the community discussed. The linguistic contrast between German and Hungarian is shown to representthe social dichotomy between a newly availableworkerstatus and traditionalpeasantstatus. December I975. In accordance with the sociolinguistic assumption that speech differences reflect the social distinctions deemed important by the community of speakers. Berkeley. the present study suggests that women's speech choices must be explained within the context of their social position. their strategiclife choices and the symbolic values of the availablelinguistic alternatives(language and sex roles. i-i6. An earlier version of this paper was presented at the symposium on 'Language and Sex Roles' at the 74th Annual Meeting of the AAA. Printed in Great Britain Peasant men can't get wives: language change and sex roles in a bilingual community' SUSAN GAL Universityof California. as well as in [I] The data reported here were gathered during 1974 as part of dissertation fieldwork supported by a N. Young women are further along in the direction of this change than older people and young men. universally.Lang. 7. A. Soc. thus the choice of languagein interactionis part of a speaker's presentation of self. in syntax and pragmatics (Keenan 197I. Bodine I975).that is.I. European bilingualism). Young women's stated preferences concerningthis social dichotomy and their changingmarriagestrategiesindicate that their greateruse of German in interactionis one aspect of their general preference for the worker's way of life it symbolizes. John Gumperz and E. INTRODUCTION Differences between men's and women's speech are no longer thought to be characteristic only of 'exotic' languages and need no longer be categorical differencesin order to be noticed by linguists (cf. interactionalanalysis.

one of the languages has come to symbolize a newly available social status. can convincingly explain why women are linguisticallyinnovative in some communities and not in others (Nichols I976). This is one of the patterns which has been noted in correlationalstudies of phonologicalchange in urban areas. nor Labov's (1972) allusion which allowwomena widerexpressiverange to normsof linguisticappropriateness than men. where it has been demonstratedthat. In other cases women do not lead in the course of linguistic change (reported in Labov 1972). In the languageusage patterns to be described here. men's and women's ways of speaking are viewed as the results of strategic and socially meaningful linguistic choices which systematicallylink languagechange to social change: linguistic innovation is a function of speakers' differentialinvolvement in. . in the linguistic repertoire of the bilingual community to be described here. . while moving furthertowards prestige models in formal speech. In the presentstudy. young women are more advanced or further along in the direction of the linguistic change than older people and young men. Specifically. social change. The substantive aim of this paperis to describethe way in which the women of a Hungarian-German bilingual town in Austria have contributed to a change in patterns of language choice. adequate explanationsof them have not been offered. use more of the new non-prestigious forms in casual speech. The young wornen 2 . Although such findings are well documented. In many cases women. Women's role in langluage change has rarelybeen linked to the social position of women in the communities studied and to the related question of what women want to express about themselves in speech. as comparedto men of the same social class.SUSAN GAL interaction(Zimmerman& West 1975. The entire community is gradually and systematically changing from stable bilingualismto the use of only one languagein all interactions. the sexual differentiationof speech often plays a majorrole in the mechanism of linguistic evolution' (Labov 1972: 303). advanced forms more frequently than men. However.Sex-linked differences in language choice have influenced the overall community-wide process of change. the effects of such sex differenceson linguistic changehave so far been noted only with respectto phonology. Most such stuldies report that women use the newer. Newly introduced forms used mostly by women are sometimes prestigious (Trudgill 1972) and sometimes not (Fasold I968). '. bibliographyin patternsof conversational Thorne& Henley1975). Young women's language choices can be understood as part of their expressionof preferencefor this newer social identity. and evaluationof. along with other social correlatesof synchronic linguistic diversity such as class and ethnicity. General statements about the linguistic innovativeness or conservatismof women will not account for the data. Neither Trudgill's (I 972) suggestion that women are 'linguisticallyinsecure'.

Franklin I969). Since World War II. These 'dialects' are not homogeneous.bilingualismin Germanand Hungarianbecamecommon. In order to make this argument in detail severalwords of backgroundare necessary. The town itself has been a speech island since the I5SOOS when most of the original Hungarian-speaking populationof the regionwas decimatedby the Turkish wars and was replaced by German-speaking(and in some areas Croatian-speaking) settlers. During the last thirty years Oberwarthas grown from a village of 600 to a town of over 5.000 people because.LANGUAGE CHANGE AND SEX ROLES IN A BILINGUAL COMMUNITY of the community are more willing to participatein social change and in the linguistic change which symbolizes it because they are less committed than the men to the traditionallymale-dominatedsystem of subsistence agricultureand because they have more to gain than men in embracing the newly available statuses of worker and worker'swife.f. Ervin-Tripp 1972). but rather are best characterizedas sets of covarying linguistic variables which have their own appropriatesocial uses and connotations(c. In Oberwart.g.The bilingual community today constitutes about a fourth of the town's population. THE COMMUNITY Oberwart(Felso'o'r) a town located in the province of Burgenlandin eastern is Austria. It is possible for 3 . Oberwartis an example of the familiarpost-warprocess of urbanization and industrialization the countrysideoften reportedin the literatureon the of of transformation peasant Europe (e. who have been trained in commerce or administration. In short.By 1972 only about one third of the bilingual populationwas employed exclusively in peasant agriculture. The indigenous bilinguals who will be the focus of this discussion have until recently engaged in subsistence peasant agriculture. It has belonged to Austria only since 1921 when as part of the postWorld War I peace agreementsthe province was detached from Hungary. THE LINGUISTIC REPERTOIRE Bilingual communities provide a particularly salient case of the linguistic heterogeneity which characterizesall communities. In Oberwart the linguistic alternatives available to speakers include not only two easily distinguishable languages but also dialectal differences within each language. have become industrialworkersor workerhowever. it has attractedmigrants. mainly people from neighboring villages.which was the largest of the five remaining Hungarianspeakingcommunities. as the county seat and new commercial center. Gumperz I964. These new settlers have all been monolingual German speakers.first about the communityand second about its linguistic repertoire. invariantstructures. most of the agriculturalists peasants.

and even young people are considered bilingual. Oberwartersconsider themselves Austrians. The young workers know that they themselves sometimes speak Hungarianand they can report on their language choices accurately. It will be arguedhere that because codes (in this case languages)are associatedwith social statuses and activities. at least for young people. and although the choice between local and standard features in either language carries meaning in conversation. The world of work is a totally German-speakingworld. changes in languagechoice can be used by speakers to symbolize changes in their own social status or in their attitudes towards the activities the languages symbolize. not Germans. It has been pointed out that a speaker'schoice of code in a particularsituation is part of that speaker's linguistic presentation of self. The choice. THE MEANINGS OF CODES Although bilingual Oberwartersuse both standardand local varieties of German as well as of Hungarian. Gal I976: III). not peasant.2This is not a reference to citizenship. nor to linguistic abilities. Of the many functions that code choice has been shown to serve in interaction (Hymes I967) this paper focuses on just one and on how it is involved in change. The speakermakesthe choice as part of a verbal strategyto identify herself or himself with the social categories and activities the code symbolizes. All old peasants do speak Hungarian and speak it in more situations than anyone else. There is no contradiction here. often using Hungarian in interactions with elders.that only the old peasants speak Hungarian. here we will be concerned only with the symbolically more important alternation between German of any sort (G) and Hungarianof any sort (H). then. The peasant parents of young workers often say about their children '0 ma egisz nimet' (He/she is totally German already). The preferred status for young people is worker. Today in Oberwart H symbolizes peasant status and is deprecated because peasant status itself is no longer respected. The phrase [2] The orthography is a modified version of Imre (I97i) and of the Hungarian dialect atlas. alternate codes within a linguistic repertoire are usually each associated with sub-groups in the community and with certainactivities.SUSAN GAL bilingual Oberwartersto move along a continuum from more standardto more local speech in either of their languages (cf.The saying refers not to actual practice but to the association of the Hungarian language with peasant status. 4 . Young bilingual workers often say. allows the speakerto express solidaritywith that category or group of people. As Blom & Gumperz (I972) have argued. Peasant is used here for a native cultural category that includes all local agriculturalistsand carries a negative connotation. and the language itself has come to represent the worker. in Hungarian.

for various reasons.G.LANGUAGE CHANGE AND SEX ROLES IN A BILINGUAL COMMUNITY indicates the strong symbolic relationshipbetween the young people's status as workersand the language which they use at work. specificationof the identity of the interlocutor was sufficient to define the social situation for the purposes of the present analysis. as in Tables i and 2. passing as a monolingual German speakeris now the aim of young bilingual Oberwarters.Therefore. regardless of which parent is bilingual. German also represents the money and prestige available to those who are employed. In predicting an individual'schoice between the three possibilities . Such general statements about symbolic associations between languages. the three factors which must be known in order to predict choices and to describe the changes in these choices are the speaker's age and sex and the nature of the social network in which that speakerhabituallyinteracts. Gumperz (1976) has called this conversational code-switching. That is. Note that 5 . A descriptionof languagechoice in such situationsmust include such variationand in this sense is comparableto the rule conflicts describedfor syntactic change by Bickerton (I1973). H or both . purposeor occasionwere largelyirrelevant. Other aspects of the situation such as locale. while in previousgenerations the ability simply to speak both German and Hungarianwas the goal of there is a premium not just on speakingGerman. While in most situationsone or the other languageis chosen. For instance. they choose to present themselves as peasants. German thereforecarriesmore prestige than Hungarian. We can think of informants as being ranked along a vertical axis and social situationsbeing arrangedalong a horizontalaxis.the habitual role-relationship between participants in the interaction proved to be the most important factor.but on speakingit withoutany interferencefromHungarian. there are some interactions in which both appear to be equally appropriate.When both languagesmayappropriately be used Oberwarterssay they are speaking 'ehodzsan dzsun' (as it comes). In addition. HOW DO LANGUAGE CHOICE PATTERNS CHANGE? In any interactionbetween bilingualOberwarters choice must be made between a G and H. but not availableto peasants. Besides the values associated with languages. In such interactions it is impossible to predict which language will be used by which speakerand both are often used within one short exchange. although H is negatively evaluated by young people it is nevertheless used by them in a number of interactions where.The children of a monolingualGermanspeakerand a bilingual speaker never learn Hungarian. Parents often boast that in their children's German speech 'Nem vag bele e madzsar'(The Hungariandoesn't cut into it). social statuses and the evaluationsof those statuses do not in themselves predict language choice in particularsituations.

Similar scales based on systematic observation of language choice were also constructed. the questionnaire results were corroborated by direct observation of language choice. for women go9%). That is. For almost all speakers 6 . M N 0 P Q R I5 H 25 H 27 H I7 H 39 H 23 H 4o H 52 H 4o H 35 H 6i H 5o H 6o H 54 H 63 H 64 H 59 H GH GH GH H H H H H H H H H H H H H H H GH G G H H H H H H H H GH GH GH H H H H H H H H GH G G GH GH GH H H H H H H G GH GH GH H H G G G G G G G G G G G G G G H H H H No.SUSAN GAL all speakers listed in these tables are bilingual. There was a high degree of agreement between observed usage and the questionnaire results (average agreement for men 86%. Language choice pattern of women Social situations (identity of participant) 8 9 6 7 5 4 G G GH GH H H H H GH H H H H H H H H H G G G G GH GH GH GH H H H H H H H H H H G G G G G GH H H H H H H H H H H G G G G G G G G G GH H H H H H H H H Infor- Age I I4 2 mant A B C D E F G H I 3 10 II H J K L.Hungarian. Note that for all speakers there is at least one situation in which they use only H.German. of informants = i8 I = 2 = Scalability = 95. TAB L E i.both German and Hungarian. The language choices of a particular informant in all situations are indicated in the rows of Tables i and 2 and the choices of all informants in a particular situation are indicated in the columns. form a nearly perfect implicational scale. arranged in this way. The choices of Oberwarters.404 7 = spouse 8 = children and their generation 9 = bilingual government officials io = grandchildren and their generation i I= doctors to god grandparents and their generation bilingual clients in black market parents and their generation friends and age-mate neighbors brothers and sisters 3 4 5 6 = = = = G . The information is drawn from a language usage questionnaire which was constructed on the basis of native categories of interlocutors and linguistic resources. H . GH .

there are some situations in which they use both G and H and some in which they use only G. When one speaker'schoice of language in a particularsituation is known it also gives 7 . if G is used with an interlocutorthen only G is used to interlocutors listed to the right of that.Hungarian. Further. H H G H H GH GH H H G G H H No. The presence of any one of the three linguistic categoriesin a cell restrictswhich of the three may occur in the cells above and below that one. then all others use H in that situation as well. with whom the speaker uses H. and GH or H are used with those listed to the left. In addition. and consideringnot one speakerat a time but the group of speakersas a whole.2%. 7 = spouse 8 = children and their generation 9 = bilingual government officials = grandchildren and their generation IOi = doctors GH . for any speakerthere are no bilingual interlocutors with whom she or he speaks both G and H unless there are some.LANGUAGE CHANGE AND SEX ROLES IN A BILINGUAL COMMUNITY T AB L E 2z.both German and Hungarian. listed to the left of that interlocutor. looking at the columns instead of the rows in Tables I and 2. of informants= x= 2= Scalability to god grandparents and their generation 3 bilingual clients in black market mko 4 = parents and their generation = friends and age-mate neighbors 6 = brothers and sisters G . then speakerslower down can be expected to use H or both in that situation. The occurrence of any of the three linguistic categories in a cell implies the occurrence of particularothers in the cells to the left and right.German. With few exceptions. Languagechoice pattern of men Informant A B C D E F 6EHH G H I Age I 2 3 Social situation (identity of participant) 6 8 4 5 7 9 G GH GH H H H H H H G G G GH GH H H H H G G G G GH H H H H G G G G G GH GH GH H IO Il I7 H 25 H 42 GH H H H H zo HH 22 H H 63 HH 64HH 43 H H G G GH H H H G G GH GH G G H G G G G G G G G G I K L M N 4i 54 65 74 58 HH H H H H H H G H I4 H H H H H H H H H H H H H H H H H H H H H GH H H H H 95. H . we see that if a speakerhigh on the list uses both G and H in a particularsituation. But if the speakerat the top of the list uses H.

Inapplicable cells (those left empty in Tables i and 2) were omitted from the denominator. (b) those who lived in householdswhich owned neitherpigs nor cows. Kovacs I942: 73-6) shows that present-day age differences are not due to age-gradingof language choice. in most ownstatus. All are significant at the O. Tables i and 2 have scalabilities of 95. The averageamount of time for all informants was seven days. the more H he or she uses. cases aperson's was not as accuratea predictorof his or her choicesas the statusof theperson'ssocial contacts.400 and 95. The closer an informant is to the top of the list the more situations there are in which he or she uses G.2% respectively. Imre I973. but ratherwas chosen to represent the entire rangeof the two variables. age. The three-way relationshipbetween languagechoices. we can take age (apparent time) as a surrogatefor repeatedsamplingover realtime ( that conclusions could be drawnabout the effect of each variableon changinglanguagechoices.SUSAN GAL informationabout the possibilitiesopen to those lower on the list and those higher on the list. The closer to the bottom.OI level. Labov 1973). In fact.age and social network. in effect.The peasantnessof a person's network. The more peasants the individual has in her or his social network the greater the number of social situations in which that individual uses H. Table 3 shows the correlations for this sample of informants.3 Given this. Oberwarters themselves define those who own cows and pigs as peasants. Two factors determinethe degree to which a person uses H as opposed to G: the person's age and her or his social network. it is worth considering the factors that determine the place of a speakeron the scale. Social network is defined here as all the people (contacts) an individual spoke to in the course of a unit of time. and peasantnessof social network can be demonstrated by ranking informants on each of the measures and then correlating the rankings with each other. Each of these network contacts was assigned to one of two categories: (a) those who lived in households which owned either pigs or cows.These results lend support to the notion that social networksare instrumental in constrainingspeakers' linguistic presentation of self (Gumperz i964. Labov I965 for details of this strategy). expressed as the percentage of contacts who fit into category (a) is. In order to distinguish the effects on language choice of time on the one hand and the effects of changing social networks on the other. [3] 'Scalability' is the proportion of cells that fit the scale model.workeror some gradationin between.whether peasant. 8 . both old people who had never been totally involved in peasant agricultureand young people who were very much involved were included in the sample. a measureof that person's social involvementwith the categoryof persons with which the use of H is associated. showing that there are only a few exceptions to these generalizations.Becausehistorical evidence (cf. Note that this group of informants was not formally selected as a representative sample of the bilingual community.

summarized in Table 3. The rule for one situation is always first categoricalfor the old form (H).it is possible to hypothesize the following process of change.74 Men o.78 Language choice and age Language choice and peasantness of network o.78 and o. in a parallel but separate process. before it is categoricalfor the new form (G).69 respectively). In fact they are not: for men the correlationbetween social network and language choice is about the same as the correlationbetween age and language choice (0. as time passes new generations use H in fewer and fewer situations regardlessof the content of their social networks.93 0. the rank correlations of language choice. the more peasants in one's social network the more likely it is that one will use H in a large number of situations. However. Young people who interactonly with workers use the least H.82 0. Both men and women show the same kinds of implicational relationships in the same ordered list of situations.The older one is the more likely it is that one will use H in a large number of situations. For women age alone is more 9 . age and peasantness of network. Correlations betweenlanguagechoiceand age. Here the issue is whetherage and social networksare equally well correlatedwith languagechoice for men and women.oi level. As speakers'networksbecome less and less peasantthey use H in fewer and fewer situations.69 0. On the basis of the rankcorrelationsthe following brief outline of the synchronic patternof languagechoice can be drawn.LANGUAGE TABLE CHANGE AND SEX ROLES IN A BILINGUAL COMMUNITY 3 . present a more complicated picture. Becausehistoricalevidence rules out the possibilityof age-gradingand because the sample allows one to disentanglethe effects of time and that of networks. Changes in language choices occur situation by situation. languagechoice and peasantness network* of All informants Women 0. DIFFERENCES BETWEEN MEN AND WOMEN The implicational scales describing choices seem to indicate no differences between men and women. while very young people who associated mostly with peasants use more H than others their own age.78 N= * 32 N= r8 N= 14 Spearman rank correlation coefficients all significant at the o. older people who interactmostly with peasantsuse the most H. And. For the sample as a whole. Older people who associate mostly with workers are closest in their language choices to people much younger than themselves. then variable(GH).

separate the men from the women and those with very peasant networks from those with nonpeasant networks.B 11l 14-34 35-55 N=14 Age Groups WOMEN -0 a)100% cL 0) 0 74 73 XL ?48 (0 ji' '0 _ _ _ _ _ _ _ _ _ _ _ _ 10 _ _ I 14-34 II 35-55 III 56-76 N=18 m Age Groups Informants with Peasant Networks Informants with Non-peasant Networks E F I GU RE i.SUSAN GAL closely correlated with language choice (0. In short there is a difference between men and women in the way each is going through the process of change in language choice. 2 56-76 0 S+. it is possible to illustrate the process at work.74).05 level. This difference between men and women is significant at the 0. Informants' networks ranged from i3% peasant contacts to MEN 100% a) CD o 60 c ~~~~~~~~55. If we distinguish three twenty-year generations. Percentage of G and GH Language Choices of Informants with Peasant and Nonpeasant Social Networks in Three Age Groups I0 .93) than is the social network measure (0.

for these women. they reject peasant life as a viable alternative. Young women with peasant networks use Hungarian as rarely as young women with nonpeasant networks. Recall that for all the men. peasantnessof networkdid makea differencesince it was associatedwith more use of H. This continuum was divided into two parts. Some of these young men have the newly developingattitudethat farmingcan be an occupation. The youngest generation of women differs both from the older women and from the men. including the youngest. since World War II.includingthe youngest men. To understandthese differencesit is necessaryto go back to the activitiesfrom which the languagesderive their meanings and evaluations. peasantness of social network makes no difference in language choice. but for each generationthis increase is greatest for those whose social networks include a majorityof nonpeasants. despite a general preference for industrial and commercial employment. these youngest women use more G and less H than anyone else in the community. The life of a II .Among the men the youngest group as a whole uses less H than any of the others. There are some young men who. a 'Beruf'. For women the process is different. Now that other opportunitiesare open to these young women. In work in inns. want to take over family farms. As with the men.For the most recent generation of women. open to women before World War II. have not been willing to take over the family farmwhen this opportunityis offeredto them. all those scoring below the median were in the nonpeasant network category. Figure i illustratesthe fact that for men there is a very regularpattern in the correlations.they specifically state that they do not want to marry peasant men. These are men whose familiesown enough land to make agricultureif not as lucrative as wage work at least a satisfactorylivelihood. First. and thereforeof social contacts. First we find that in the oldest generation this sample includes not one person with a nonpeasant network. young women. Regardlessof the negative evaluations. But those young men with heavily peasant networksdo use more H. if only temporarily. like any other. From the oldest to the youngest generation use of G increases. This is not a sampling errorbut reflects the limited range of activities. All those scoring at or above the median were put in the peasant network category in Figure I. those who have heavily peasant networks use more H than those who do not. More importantly. It will be argued here that their language choices are part of this rejection.for these young men expression of peasant identity is still preferredfor many situations. factoriesand shops. Often they remainedin contact with those they befriended. In contrast. Many women of the generation reachingmaturityduring and after World War II left the peasant home.LANGUAGE CHANGE AND SEX ROLES IN A BILINGUAL COMMUNITY 94% peasant contacts. In the middle generation the women's pattern matches that of men exactly. peasant life is a much less attractivechoice than it is for men.

nevertheless for the peasant wife the independence which is said to compensatethe peasant man for his work is not freely available. But in peasant households the male labor saving machines are always acquiredbefore any of the ones which lighten women's work. where. is never built before a combine is purchased.and the combine itself is among the last and most expensive of the machines acquired. machines such as the tractor and the combine make men's farm work highly mechanizedand considerablyless difficult than it once was. and (c) to be a full-time agriculturalist. being a young peasant wife often means living under the authority of a mother-in-law who supervises the kitchen and women's farm work generally. Because of the increasedaccess to money. The peasant wife typically spends the day doing farm work: milking. But it is generallyagreedthat while agriculturalwork was once more grueling and difficult than factory and contruction work.but mostly they hold part-timejobs or are not employed at all. In addition. in the Balkans(compare Denich 1975 with Fel & Hofer I969: I I3-44). For women the life possibilities depend mainly on whom they marry. For instance the silo. the German small peasant'swife in I964 averagedover the year seventeen more work hours per week than her husband (Franklin I969: 37-44). the electric stove and the silo (which eliminates the need for rootcrops as cattle feed). planting and harvestingpotatoes and a few other rootcrops. In fact. Let us compare the choices open to Oberwartyoung men and women as they see them.This last is felt by Oberwartmen to have the advantageof independence . including the washing machine. marriageto a worker involves only household tasks and upkeep of a kitchen garden. hoeing. which is perhaps the most substantialwork saver for the peasantwife. for instance. for instance.holding two full-time jobs.SUSAN GAL peasant wife is seen by Oberwart young women as particularly demeaning and difficult when compared to the other choices which have recently become availableto them. Wives of workersare sometimes employed as maids or salespersons. (b) to be a orders from strangers. electric stoves and washing machines are among the first appliances bought by workingmarriedcouples. Although peasant men still work longer hours than those in industry. feeding pigs. There are machines now available which lighten the work of the peasant wife considerably. Peasant I2 . because agriculturalequipment is not needed and because some of the women themselves contributepart of the money. although peasant life in Oberwart is less male-dominated than. In this Oberwart exemplifies the pattern all over Europe.Her evenings are spent doing housework. this is no longer the case. Industriousness is traditionallyconsidered a young peasant wife's most valuable quality. thereby furtherlightening the wife's work load. For men the life possibilities are (a) to be an industrialor construction worker(usually a commutercoming home only on weekends). In marked contrast.and the disadvantage of lack of cash and prestige.

between '3 . in referenceto a particularyoung couple an old Abbu ma' man remarked:'Az e Truumfba j'ar. they do not present themselves as peasants in speech. Peasantmen workmore hoursthan worker men. If in recent years Oberwartyoung women have not wanted to marry peasant men.that's for sure. Oberwartersagree that 'Parasztlegin nem kap nuot' (Peasant lads can't get women).) Although the young men themselvesare usually also reluctantto become young women. Worker'swives. This contrast is not lost on young Oberwart women. then Oberwart peasant men must have found wives elsewhere. the anti-peasant attitudes of the community's young women present a problem.she's going to shovel cow manurefor him?She'll neverbe a peasant. When discussing life choices they especially dwell on the dirtyness and heaviness of peasant work. az fog neki tehen szart lapaini? parasztnem lesz. Mothers of marriageabledaughters specifically advise them against marriage to peasants. City of Oberwart. az ma'zicher!' (She works at the [local bra factory]. often work fewer hours than their husbands.Between I96I and 1972 only 32% were. of peasantstatus and life generally. Endogamous marriagesof all bilingual Oberwarters bilingualmalepeasant Oberwarters and % Endogamous marriages of all marriages 1911-40 194i-60 % Endogamous marriages of male peasants 87/o 7I% 65 32 54 0 I96I-72 Source: Marriage Register. For instance.for those who nevertheless choose family agricultureas their livelihood. They do not want to be peasants. For instance. But for the bilingual peasantmen of Oberwartthe figuresare different. The town's marriagerecords should provide evidence for the differencein attitudes between young men and young women.LANGUAGE CHANGE AND SEX ROLES IN A BILINGUAL COMMUNITY wives workfar morethan peasantmen. Rejection of the use of local Hungarian. Table 4 shows that between I9II and 1940 7200 of the marriages of bilinguals in Oberwart were endogamous. can be seen as part of the rejection. and if they have acted on this preference.As Table 4 indicates. TABLE 4. especially if not employed. The general trend in Oberwartin the post-war years has been away from the traditionalvillage endogamyand towards exogamy. the symbol of peasant status.

Because the children of marriages between monolingual German speakers and bilingual Hungarian-German speakers in Oberwart rarely if ever learn Hungarian. women have less to lose in rejecting the traditional peasant roles and values. evidence was presented showing that in their stated attitudes and their marriage choices the women evaluate peasant life more negatively than the men and reject the social identity of peasant wife. that is the ones engaged in peasant agriculture.SUSAN GAL I9II and I940 a larger percentage of peasant men married endogamously than all bilingual Oberwarters (87%). They refuse to marry local peasant men. This paper has argued that women's language choices and their linguistic innovativeness in this community are the linguistic expressions of women's I4 . In exploring the reasons for the difference between young men's and young women's language choices. by I96I-72. Finally. CONCLUSION There are two ways. Indirectly. the young women. The effect of this is discussed below. however. Those peasant men who did marry during those years found wives in the neighbouring small German monolingual villages where being a peasant wife has not been negatively valued. in most situations. In short. this was reversed. The women of Oberwart feel they have more to gain than men by embracing the new opportunities of industrial employment. Also. As a result. Directly. Between I94I and I960. to present themselves as peasants by using H. preferring workers instead. when 32% of all bilingual Oberwarters married endogamously. considering the male-dominated nature of East European peasant communities generally and the lives of Oberwart women in particular. in an indirect way the present generation of young women is limiting the language possibilities of the next generation. exactly that small group of young men most likely to be using Hungarian in many situations. It also contrasts with the choices of young men. not one peasant man married endogamously. even those with heavily peasant networks refuse. who use Hungarian in more interactions than the young women and who are constrained by the peasantness of their social networks so that those with heavily peasant networks choose local Hungarian in more interactions than those with nonpeasant networks. are the ones who have married German monolingual women with the greatest frequently in the last decade or so. in which the attitudes and choices of young bilingual women are changing the language usage pattern in this community. one direct and one indirect. young women's marriage preferences are also having a linguistic effect. This contrasts with the language choices of older women and has the general effect that more German is used in more interactions in the community. the marriage records provide evidence that young Oberwart women's stated attitudes towards peasant men have been translated into action.

Henley (eds). Language 44.Nipi Tdrsadalom (Folk Culture . Lamphere (eds) Woman. Berkeley: University of California. Bickerton. J. T. In S. Nichols. Gumperz. Explorations in the ethnography of speaking. Stanford: Stanford University Press. (1970). Hymes (eds). (1976). Rubin. S. J. D. 8i-iI5. and Hofer. American Anthropologist 66 (6). B. J. LinS 2. M. Norm-makers. Rinehart & Winston. The nature of a creole continuum. The linguistic consequences of being a lame. J. Fishman (ed. Sherzer (eds). Thorne & N. Black women in the rural south: Conservative and innovative. (1974). (1974). Fasold. Proper peasants. Linguistic and social interaction in two communities. In J. (1976). Verbal strategies in multilingual communication. & Gumperz. Mimeo. (I969). Sex and power in the Balkans. Gal. Directions in sociolinguistics. Budapest: Sylvester-Nyomda. Bauman & J. Rinehart & Winston. In Nipi Kultura . Kovacs. D. Denich. R. dissertation. 8-28. In M. Rowley. H. (Ic973). Ervin-Tripp. 45-80. Directions in sociolinguistics. The sociolinguistic significance of conversational code-switching. R. A. E. (1969) Speech variation among Negros in a small southern community. Hymes. Bodine. J. The European peasantry. Rosaldo & L. norm-breakers: Uses of speech by men and women in a Malagasy community. W. F61. J (I974) Social meaning in linguistic structures: Codeswitching in Norway. (I972). Philadelphia: University of Pennsylvania Labov. J. Franklin. London: Methuen. Readings in the sociology of language. (I976). (I973). Las Cruces. Guatemala. Language and women's place. Paper presented to the conference on the Sociology of the Languages of American Women.Folk Society). In J. (I968). (I964). The linguistic pattern is best understood by considering the social meanings of the available languages and the strategic choices and evaluationswhich men and women make concerning the ways of life symbolized by those languages. The Hague: Mouton. S. Budapest: Akaddmiai Kiad6.LANGUAGE CHANGE AND SEX ROLES IN A BILINGUAL COMMUNITY greaterparticipationin social change. Keenen. Massachusetts: Newbury House. Ms. (x967). Farber. New York: Holt. Budapest. Sociolinguistic patterns. J. New Mexico. Bilingual usage in Paraguay. (1970). Unpublished Ph. LinS 2.C.D. Nyelvtudomdnyi Ertekezdsek 72. (1973). (x969).). Imre. S. Women's language use in Comalapa. Blom. Chicago: Aldine. P. A. A felsoori nyelvjiris (The Oberwart dialect). REFERENCES Anshen. Lakoff. Washington. D. Journal of Social Issues 23 (2). Paper presented at the 73rd AAA Meetings. (I975). Gumperz & D. Hymes (eds). F. In J. Michigan: University Microfilms. Language and sex: Difference and dominance. E. Part II.: Center for Applied Linguistics. New York: Holt. In R. 137-54. Language Behavior Research Laboratory Working Paper * 36. London: Cambridge University Press. New York University. Gumperz & D. I5 . group in Austria). A felsoori magyar nepsziget (The Hungarian folk-island of Oberwart). culture and society. (1974). Models of the interaction of language and social setting. Mexico City. (I972). Language change and its social determinants in a biblingual community. (I97x). P. (1942). S. S. On sociolinguistic rules: Alternation and co-occurrence. Az ausztriai (burgenlandi) magyar szorvanyok (The Hungarian minority (1973). 64o-69. Press. A sociolinguistic study of the pronunciation of three vowels in Detroit speech. Sex differentiation in language. Ann Arbor.

179-95. & Henley. Rowley. In B. N. LinS I. Linguistic diversity in Canadian society. B. Thorne and N. Henley (eds). (I97I). & Cedergren. covert prestige and linguistic change in the urban British English of Norwich. Massachusetts: Newbury House Publishers.). (1975). Language and sex: Difference and dominance. Zimmerman. P. Inc. Some results of a sociolinguistic study of Montreal French. Thorne & N. Henley (eds). (eds). Inc. Trudgill. Sex roles. Massachusetts: Newbury House Publishers. M. Language and sex: Difference and dominance.SUSAN GAL Sankoff. (1975). Massachusetts: Newbury House Publishers. C. Swacker. (1975). D. In R. Rowley. Language and sex: Difference and dominance. (1972). The sex of the speaker as a sociolinguistic variable. Edmonton: Linguistic Research. Inc. G. Thorne. I6 . H. Rowley. & West. In B. Sex. interruptions and silences in conversation. Darnell (ed.