Introduction

This chapter considers the Hungarian lexicon of Hungarian–German bilingual speakers in an Austrian community that has been shifting away from Hungarian for about forty years. I will use evidence about the formation and use of verbs in order to address the following questions: Which word-formation devices or patterns are retained and which are lost by those speakers who use Hungarian only in a narrow range of contexts? What accounts for the fact that, despite diminishing social function for the language, narrow-users are in one sense more innovative than broad-users in several kinds of word formation?

Among documented cases of community-wide language loss there is considerable variation in the rapidity of loss and in the extent of discontinuity between generations or social groups of speakers. When, as in the present case, the shift takes more than one generation and occurs through a restriction in the social functions of the obsolescent language, it raises theoretical questions about the relationship of language use to linguistic structure (Hymes 1974). The explanation for documented linguistic differences between those who use a language in a wide range of contexts and those who do not has most often been sought in the early and continuing effects of those very differences in range of use. According to this hypothesis, the decreasing use of the obsolescent language by adults, and its increasing functional specialization, lead to the incomplete acquisition of the language by children. We can thus understand the frequent loss of linguistic structures that are not available at all for input to children.

Introduction

Studies in language obsolescence have broached significant directions of inquiry in the dynamics of language change. In scholars' efforts to circumscribe the field of language death research, the following major areas of findings seem to be the most conspicuous. (1) It is probable that language death does not differ in kind from other types of linguistic change, but in the speed with which structural changes occur and in the number of phenomena covered by the process (Dorian 1981; Schmidt 1985c). (2) In cases of contracting languages it is possible to end up with communities characterized by marked asymmetries in the development of pairs of skills such as phonology vs. grammar, passive vs. active competence, high vs. low or colloquial stylistic level, written vs. spoken discourse, which normally co-occur in a more balanced way in “healthy” mother-tongue situations (Dorian n. d.). (3) Language death studies raise crucial questions concerning the concept of the speech community, since the frequently appearing sociological category of imperfect speakers involves problems of successful participation in the various communicative events of the community (Dorian 1982a). (4) There is evidence to suggest that, despite certain similarities between language death and pidgins and Creoles, the two processes differ crucially (Dorian 1981; Schmidt 1985a). (5) Significant structural restrictions in grammar have been convincingly correlated with reduction in speech genres which were once highly valued specimens of verbal art and competence (Hill 1978; Tsitsipis 1984). (6) Linguistic competence exists among fluent speakers in the ordinary sense of productivity, but the language has stopped being a source of continuous invention (Hymes 1984).

Community profile

French settlement

The city of Welland, Ontario is located in the heart of the Niagara Peninsula close to the famous falls of the same name. The story of its French community is that of practically all communities of the French-Canadian diaspora. The exodus of French Canadians beyond the borders of the province of Quebec (the principal stronghold of the French language in North America) was essentially touched off by rural overpopulation, forcing many to leave their old settled parishes in search of manual work in the developing urban industrial centers of Ontario to the west and New England to the south. Although the initial migration of French Canadians from Quebec to Ontario dates back to the early nineteenth century, the origins of Welland's French-speaking minority are more recent, since they can only be traced back to the turn of this century. Welland's industries (chiefly textile mills, iron and steel factories, and rubber plants) were spurred by the outbreak of World War I, creating many new, well-paid jobs which were to attract, by 1919, around forty French-Canadian families from Quebec. These families may be considered the historical kernel of the French-speaking community of Welland. They settled in the eastern half of the city where the industrial plants were located, an area which was soon to be known as “Frenchtown”. By the end of the 1940s, however, social mobility had brought about some residential dispersion.

As an industrial center, Welland was hit hard by the great Depression, which saw the departure of numerous francophone families.

Introduction

Eastern Africa does not constitute a unit linguistically or culturally. Instead, this area, roughly comprising the countries of Djibouti, Ethiopia, Kenya, Somalia, Sudan, Tanzania and Uganda, is characterized by a large variety of language families and groups on the one hand, and a considerable degree of cultural diversity on the other. In one sense the present geographical delimitation for the discussion of language obsolescence is therefore arbitrary. Such a regional approach, however, allows for a comparison of different circumstances relevant to the actuation problem of dying languages. One such variable concerns the social and economic structure of interacting groups, whilst a further variable relates to demographic factors.

The present contribution is directed towards disseminating general information on the causes and effects of language shift. It is not a survey of all known cases of language contraction and death in eastern Africa. Instances of partial and complete language shift are so numerous that it can be stated without exaggeration that most ethnic groups in this part of Africa, and probably elsewhere, are the result of an amalgamation of various groups which were distinct at an earlier stage. More specifically, this paper tries to provide answers, at least in part, to the following questions:

1. To what extent, and in what ways, do economic and social factors play a role as causal mechanisms in the gradual “default assignment” of certain languages?

2. How important is language as a potential symbol of ethnic identity relative to other symbols?

3. What kind of traces, linguistically or otherwise, does language shift leave behind, once the extinction of a particular language has become a fact?

Linguistic researchers who find themselves working with severely contracting speaker groups may begin with entirely different interests and objectives, yet very likely soon find themselves confronted by intriguing linguistic and sociolinguistic phenomena which turn up in the process of their investigations.

In theory, if the group is of sufficient size at the start of the investigation and the contraction of the linguistic community is gradual, there may be time and evidence enough to identify some of the phenomena early on and to observe their development. If the group is already small, its membership not entirely clearcut, and its further contraction rapid, then the researcher's observations are likely to be more tantalizing and suggestive than full and conclusive, and the fieldwork and research problems will be compounded. The latter is the case with many diminishing Native American languages, including the one reported on here.

In this chapter I will describe several different sociolinguistic and structural aspects of Gros Ventre obsolescence. In passing, I will also refer to various problems which had to be confronted in the course of the research.

What I know about Gros Ventre obsolescence has been discovered largely by accident, and comes from random aberrant forms which have turned up in my fieldnotes, from remarks made by my informants about the speech of others, and from my own observations of life and personal relations on the reservation. I must admit, however, that some of what I claim rests on rather slim evidence, and may prove eventually to be erroneous or better explained in other ways.

Thinking about language death confronts the historical linguist constantly with awkward questions concerning his fundamental concepts. All the contributors to this collection are up against these questions whether or not they say so explicitly and, for that matter, whether or not they call themselves historians. In what follows we shall discuss a few of the more explicit comments.

There was at one time a set of preconceptions into which the subject would have fitted only too well: the framework of growth and decay and ultimate death, or of evolution and ultimate extinction. Those ideas are no longer virulent. Their latency, however, is another matter. Living individuals and living species with their rather well defined physical boundaries may have proved poor sources of metaphor; but metaphor of the kind that makes a language into a corporeal object of some sort will always be with us, and so, along with its blessings, will be its pitfalls.

In retrospect – that is, to the historian – what is obsolescence? Aside from instances where speakers are killed off (“extremely rare in eastern Africa” says Dimmendaal [this volume] in one of his many intriguing asides) or, under some kind of pressure, abandon their language with a will, obsolescence occurs when a population shifts from one language to another, in ways which are open to observation (as regards the present) and to surmise (as regards the past; Campbell and Muntzel, Dimmendaal). Usually, therefore, language death is preceded by bilingualism, though, of course, bilingualism does not always presage language death. English could have died in the Middle Ages but didn't.

In my study of the Norwegian language in America (Haugen 1953 [1969]) I surveyed the language usages of Norwegian immigrants in the United States from the beginnings of immigration in 1825 to the 1950s. I devoted two chapters (10 and 11) to the use and gradual disuse of the language over the more than a century in which it played a significant role in American life. I entitled the first chapter “The struggle over Norwegian” and the second “The triumph of English”. In the present article I shall (1) briefly summarize my findings in the book for the benefit of readers who may not be familiar with it; (2) present some personal data gathered in my fieldwork; and (3) provide a short account of what has happened in the intervening quarter of a century.

Norwegian America: the rise

My point of departure is an 1898 quotation from an eminent Norwegian-American educator: “Now the question no longer is: how shall we learn English so that we may share in the social life of America and partake of her benefits; the great question is: how can we preserve the language of our ancestors here in a strange environment and pass on to our descendants the treasures it contains? (Bothne 1898:828; cit. Haugen 1953:238). It is significant that this question was posed at the turn of the century. By that time Norwegians had been pouring into America for two generations, settling for the most part in the upper Midwest, from Illinois to North Dakota.

More than a half-million Norwegians had changed countries in these years, seeking opportunity in the rural Midwest.

For immigrant and minority languages in the United States, language death is an almost inevitable outcome of contact with American English. The promise of social and economic advancement proffered by mastery of English eventually overcomes the most fervent of language loyalty intentions. The number of social contexts in which speakers can use the minority language. steadily declines. Without the support of continued immigration from the language homeland, the number of fluent interlocutors gradually decreases, and eventually no social context remains in which it is appropriate to speak the minority language. Pennsylvania German, although it has enjoyed a long history in America, is no exception to this general process.

Pennsylvania Germans settled in America during colonial times in farm communities across southeastern and central Pennsylvania. Secondary settlements arose later in Ohio, Indiana, Illinois, and the Virginias. While many of these communities have long ago completely and irrevocably assimilated into mainstream American society, some have maintained their peculiarly Pennsylvania German (PG) culture and language to the present time. Although all Pennsylvania Germans share many traits and values, the PG population is not one homogeneous group. It consists of many subgroups, each having a different relationship to the dominant society, and each speaking a variety of Pennsylvania German which fulfills different communicative and symbolic functions. The purpose of this study is to document linguistic change in such subgroups: Pennsylvania German, as it is dying in farm communities of nonsectarians, and as it is being maintained in farm communities of Amish and Mennonite sectarians.

Background

Within a community shifting from one language to another, proficiency in the receding language varies widely.

Introduction

In this chapter we are concerned with structural changes in obsolescing languages attributable to the language death process. On the basis of our experience with a number of dying languages we propose hypotheses about the characteristic structural developments within the languages concerned. These hypotheses can be confirmed, or contradicted and refined, as additional understanding of dying languages accumulates from the rapidly growing sources.

The languages of our experience upon which we base our observations are set forth below, presented with information on their geographic location, genetic affiliation, and number of speakers. Speakers' linguistic ability or structural knowledge of the obsolescing languages of this list varies greatly, and many communities exhibit a proficiency continuum ranging from fully competent speakers to individuals with very little knowledge at all. For purposes of exposition only, to give an idea of the kinds of speakers in each of the situations with which we worked, we characterize speakers roughly as S for “strong” or “(nearly) fully competent” I for “imperfect”, i.e. for reasonably fluent so-called “semi-speakers” W “weak semi-speakers” with more restricted speaking competence (perhaps akin to Elmendorf's [1981] “last speakers”); and R for so-called “rememberers” who know only few words or isolated phrases (“word-inserters” may belong to this group: see Voegelin and Voegelin 1977b). Language communities with the full proficiency continuum from S to W and/or R are presented as PC.

Introduction

Reduction in the frequency of relative clauses in the usage of speakers in late stages of language death has been identified in languages of diverse genetic and typological affiliations: Cupeño and Luiseño (Hill 1973, 1978), Trinidad Bhojpuri (Durbin 1973), Tübatulabal (Voegelin and Voegelin 1977b), East Sutherland Gaelic (Dorian 1981, 1982b), and Dyirbal (Schmidt 1985c). Relative clauses, along with other complex sentence phenomena, have also been reported as absent in pidgin languages, developing only upon creolization of a pidgin (Sankoff and Brown 1976). Low frequencies of complex sentences have been identified also in working-class (as opposed to middle-class) usage in British English (Bernstein 1972), American English (Wolfram 1969), and French (Lindenfeld 1972). Van den Broeck (1977) identified the phenomenon in some speech contexts for working-class varieties of Flemish. A low rate of relativization in oral, as opposed to written, language in Western languages has also been frequently noted (see Chafe 1982). Romaine (1981) found lower rates of relativization associated with more intimate, as opposed to more formal, written styles of Middle Scots. The identification of a similar pattern of change or differentiation in such a variety of contexts challenges us to develop a unified functional explanation. Such an explanation would be new support for a widely shared intuition that the processes of change in language death are differentiated from ordinary change processes primarily by their rapidity (Dorian 1981; Schmidt 1985c).

Benveniste (1971) called relative clauses “syntactic adjectives” Foley (1980) has pointed out that they are the “most sentential” noun complements. Unlike adjectives, participles, and gerunds, relative clauses include tensed verbs.

In dealing with the similarities and differences between dying languages, pidgins, creoles, and immigrant languages, I will concentrate my remarks on two areas: the nature of the sociolinguistic context which gives rise to these varieties of language, and structural factors common to these cases. Finally, I will draw some conclusions about the implications of the study of these languages for linguistic theory and suggest some directions for further research.

In attempting to sketch a sociolinguistic profile of the context in which pidgins, creoles, dying, and immigrant languages emerge, the most general factor that can be mentioned is that of language contact. Seuren (1984:209) says that whatever is universal in creoles is also characteristic of contact languages of any kind that get turned into native languages. Even leaving aside the question of how much nativization owes to universals, at the moment there is not enough systematic descriptive and theoretical work in language contact to draw up a typology of parameters from the relevant cases. This collection is a particularly useful beginning.

I will not single out immigrant languages as a special case because immigrant languages may of course be as healthy as varieties of these languages spoken elsewhere. Typically, however, for a variety of political and social reasons, which many of the contributors here discuss in detail, they aren't. The fact that healthier varieties of these languages exist elsewhere for comparison is in many ways a methodological advantage. Another, on occasion, is the existence of a range of communities using a restricted variety for different purposes (see, for example, Huffines on Pennsylvania German, Mithun on Cayuga, also Hamp 1978 and this volume).

When a language has been in retreat for a long time and its distribution has been shrinking at the same time that its functions have been dwindling, difficulties are very likely to arise in even such basic matters as determining just who should be considered a “speaker” or a “member” of the speaker community. The “native speaker” population itself may not agree on who falls within that category: some people may claim speaker status when others would not accept them as such; some may say that they are not speakers when others would include them as speakers. If the speaker population cannot agree on its own membership, then the problems for the researcher are bound to be even more acute. This chapter tries to deal with each of these matters in turn, community structure and speaker self-identification on the one hand, and encounters between communities and the investigator in search of speakers on the other.

The community

Irish and Scottish Gaelic-speaking areas

The modern native Irish and Scottish Gaelic-speaking populations, now largely confined to the extreme northern and western fringes of Ireland and Scotland and having the appearance of being self-contained entities are, nonetheless, in many ways a community within a community. Their members for the most part share the same racial inheritance as the general Irish and Scottish populations respectively and have religious affiliations in common with a greater or smaller section of the larger community. But above all, they must struggle for a place alongside English, that giant among the world's modern languages.

Introduction

The results of intensive language contact can have the effect of accelerating internally motivated changes in the system of the less-used language, and direct influence from the majority language will be difficult to resist. The changes in the minority language will, however, not be abrupt but will proceed step-by-step. In this chapter I will discuss the changes in the objective case-marking system in Estonian spoken as a minority language in Sweden. The Estonian objective case-marking system is much more complex than the Swedish one: this gives rise to the kind of development reported for another minority language under pressure, where “a system of morphological markers is confined primarily to a single high-frequency member of the word class concerned”, and “competing structures that have the same semantic value, but different constraints, show movement toward a single favored structure” (Dorian 1983:161,160).

Sweden as a multilingual and multicultural country

Sweden has been a multicultural country for a very long time with two indigenous minorities within its borders: the Finns in the Swedish part of the Torne Valley at the Finnish border, and the Lapps, who probably were the original population of northern Scandinavia. There has also been immigration to Sweden during several periods of Swedish history (e.g. Arnstberg and Ehn 1976; Wande 1984). Some ethnic groups have maintained their culture and ethnic identity while others have been absorbed into the Swedish majority.

Estonians, Latvians, and Lithuanians came to Sweden as refugees during World War II, mainly in 1944, in total approximately 35,000 people, and they were followed by labor-market immigration mainly in the 1960s and 1970s.

Introduction

Sudden shifts in a community's linguistic usage patterns afford sociolinguists interested in “explaining” linguistic changes a tempting opportunity. In cases of gradual shift, it is easier to understand change as the result of myriad social and linguistic factors and pressures, each contributing incrementally to the final result. With sudden change, on the other hand, we are more tempted to look for the one cause or factor tipping the linguistic balance. But, even in cases where rapid shift is clearly a response to a particular social change – for example, Dorian's account of shift away from Gaelic as the Scottish Highlands became less isolated (1981:51), or Tabouret-Keller's of shift in response to ongoing industrialization (1972) – there is still a complicated web of social, linguistic, and ideological factors at work. Thus Dorian notes that “English seemed to have come quickly to eastern Sutherland, but the climate which led to its rapid adoption had been centuries in the making” (1981:51). Dorian's caveat is even more clearly a propos where there is no sudden social shift corresponding with a change in language use.

The Scottish Gaelic speakers of Cape Breton, Nova Scotia, experienced a rapid language shift during the 1930s and 1940s. Their case is an excellent example of linguistic “tip”, as Dorian first defined it: “A language which has been demographically highly stable for several centuries may experience a sudden ‘tip’, after which the demographic tide flows strongly in favor of some other language” (1981:51).

Introduction

In the light of more than two decades of research in the Labovian framework, conducted mainly in urban speech communities, there exists a widespread belief that linguistic variability is most likely to appear when there is strong socio-economic stratification, or even that it only appears under these circumstances. This is the impression left, intentionally or not, by statements such as the following by D. Sankoff and Laberge (1978:239):

and by Labov (1966:111):

Quantification of linguistic data according to linguistic environment and according to the speaker's social class is basic to the methodology of this branch of sociolinguistics. It is also an underlying assumption that variables involved in a linguistic change necessarily carry sociosymbolic meaning.

In the language death literature, it has been shown that variation may arise in languages undergoing attrition as a result of a language death process whereby simplified variants gradually replace more complex variants, especially in the speech of semi-speakers (see e.g. Dorian 1978c). Schmidt (1985a) also gives evidence for the creation of new, complex variants in dying Dyirbal.

In Old Nubia, that is, before the 1964 resettlement (see Chapter 6), the Nubian language was resistant to Arabic interference because it was an undisputed language in a remote place. Women and children were mostly monolingual Nubian speakers. Men in general spoke both Nubian and Arabic due to the fact that they had to travel to cities such as Cairo and Alexandria looking for wage labor. However, there is a noticeable difference in the trend of migration after the resettlement of 1964. Labor migration to big cities is not as high as it was prior to resettlement, due to the availability of job opportunities in New Nubia.

Nubians can be found as well-entrenched and well-assimilated urbanized groups in cities, and on the other hand as non-urbanized groups living in Southern Egypt in small villages. Groups in several types of locations were chosen for this study. The first group was drawn from isolated villages, where Nubian inhabitants must take a bus to go to the nearest city. The second group was drawn from villages near Aswan, where Nubians can walk or cross the river by boat to that city. The third and final group was drawn from the cities of Aswan, Cairo, and Alexandria, where Nubian men work, but still maintain a home in their Nubian village in Southern Egypt. The different locations were chosen because they give a representative sample of urban, quasi-urban, and non-urban Nubian speakers.

In the new community created after the resettlement of 1964, the Egyptian Nubians came into closer contact with non-Nubian Egyptians. Nubian, an East Sudanic language, came into contact with a dominant Semitic language, Arabic.

During the preparation of this volume, a participant suggested that the resulting collection might represent a “state of the art” milestone roughly comparable to Wanner and Gleitman's (1982) Language acquisition volume. Looking closely at that suggestion is perhaps a little dispiriting, but it's also revealing.

The subfield of child language is designated as foremost specialty by a fair number of linguists these days. There are at least two journals (The Journal of Child Language, First Language) devoted to the subject, and there are well-established annual conferences, in addition to special sessions, workshops, and so forth, at larger meetings. Researchers in child language have enough recognized outlets and enough predictably scheduled get-togethers so that they can identify one another and exchange findings and ideas with relative ease and speed. The impulse for this volume was the lack of those things (as yet) in the field of language obsolescence.

Few of the contributors to the first two parts of this book would identify themselves as specialists in some one particular area which this volume could be said to be devoted to. There is even some difficulty in locating any such area: “obsolescence”, “contraction”, and “death” all fall short of a precise (perhaps even an acceptable) rubric; see especially the commentaries, but various authors in Parts I and II wrestle with this problem as well. Not only is there uncertainty about the designation of the field, but in addition other terminological uncertainties hamper communication among researchers.