Identity and Language – Language Death in Africa

Nami Ando May 3, 1995

The relationship between language and group identity is a source of continuing controversy. An individual may identify himlherself as a member of multiple groups on multiple levels. This would affect the individual’s choice of language (or dialect), and ultimately the vitality of that language. Changing conceptions of identity, in turn, affect language. Languages adapt to its changing social function and context of usage. Often, however, language is analyzed as an autonomous system; the process of language change IS thought to be part of natural linguistic development. At times the change is indeed natural, i.e., not externally motivated. But language is spoken by people, and one cannot discuss language change without discussing the change among people who speak it. This is why it is useful to look at the different ways in which group identities interact, in order to get a complete picture of language change.

Language death is an extreme form of language change. At one time it was thought that languages were organic, and that they had a life of their own. According to this theory, all languages have a natural lifespan, and languages, like plants, eventually reach a stage in their development where they wither and die. The weakness of this theory is that ignores the fact that the fortunes of languages are bound up with those of its users. If languages decline or die it is simply because the circumstances of their speakers have changed. From a reading of articles written on language death in general, language death seems to fall into two broad categories: language death without language shift, and language death with language shift.

The second type is the one most commonly analyzed, and the one that will be discussed in this paper. Language death without language shift occurs when, either the speech community suddenly dies through disease or genocide, or the language is suppressed altogether through political oppression. Language death with language shift is more common. This type of language death can occur in a number of ways. For example, the speakers may disperse and take up different languages. Also, there could be an encroachment of a dominant language, so that the language gradually declines in use and is replaced by another language. Typically language death is preceded by bilingualism, although the opposite does not always hold.


Some linguists use such value-laden terms as « language murder » or « language suicide » to characterize instances of language death. It lends an air of tragedy to the process. Besides attaching a normative value to the health of languages without taking into consideration the social context and welfare of the speakers, such words are misleading because they give the illusion of an autonomous language system acting upon itself or another language system. Language shift often reflects pragmatic desires for higher social mobility on the part of the speakers. Furthermore, even the word « language death » itself is laden with several definition problems. For example, it is a matter of definition when one chooses to pronounce a language to be officially « dead »- when all semi- speakers are physically dead, or when the language has ceased to be used in any speech community? Also, it is difficult to draw the boundary between what « counts » as a language and what doesn’t. Often whole speech communities will shift from a distinct dialect to the standard form within the same language.


The symptoms of language death are often hard to separate apart from its causes. The symptoms can be divided into two obvious kinds: social and linguistic. There are several identifiable social symptoms. which do not necessarily apply to all cases. First. and most obvious. is that the number of speakers of the language is steadily decreasing. This is a disputable issue, because no one is sure exactly how many speakers are « enough » to sustain a language. Second, languages in decline characteristically have a predominance of middle-aged or elderly native speakers. This means that the language is not.being transmitted to the younger generation. Third. languages in decline are often confined to rural areas in which associations are made between the the particular language and low status. This is clearly disputable in areas where modernization is not a direct factor of language decline. Fourth, diglossia is common in areas of language decline. This is an intermediary and temporary stage, before it is replaced with dominant language monolingualism. Diglossia is defined as « the status of a language in a whole
community where two languages or very different variants of the same language coexist, e.g., a standard and a minority language with different functional domains » (Dressler and Wodak-Leodolter, 8).

Diglossia is to be distinguished from bilingualism. which is an individual phenomenon referring to any person who has learned and internalized a specific combination of languages and uses them according to general and individual principles of language choice. It can be said that when a language possesses no more monoglots. the process of decline has often already begun. The only instance where diglossia can be a fairly stable condition IS when there are well


defined and substantial domains of use for each language. A
relevant question might be to ask how long the state of diglossia In a speech community can be sustained.

In terms of its linguistic aspects, language contraction seems to involve some kind of reduction. The problem, however, is that it is often difficult to tell the difference between change that affects a language that happens to become extinct, and change that can specifically be diagnosed as leading to the extinction of the language. Usually the linguist only knows by hindsight which linguistic changes lead to language death.

Nevertheless, many types of linguistic reduction that seem to be symptomatic of language decline have been documented. One claim is that languages about to disappear suffer style reduction. (Hoenigswald, 348) This occurs during the period of bilingualism, when speech of a particular style to be used in particular contexts IS left to the other language. Thus a language may retain its linguistic variation but become restricted in its contexts of usage, thus losing a stylistic option. Another claim is that change during obsolescence may not differ from change in general, but that it proceeds faster.

This assumes a « nonnal » rate of change common to all languages, which, although it has been attempted, is inevitably a messy affair. The final, and most complicated, claim is that languages in decline go through simplification. The areas in which simplification is seen is quite diverse. There are two kinds of simplification. The first is where distinctions not known to the superstratum are given up and there is a convergence of features. The second is autonomous simplification per se, of the pidgin variety.


The first kind of simplification may involve (and the list is far from complete) loss of morphological categories, allomorphic leveling, loss of case markers, and loss of agreement markers. In Dorian’s study (1991) of East Sutherland Gaelic, for example, she found the following phenomena which she « suspect[s]…are in fact characteristic of languages in decline » (27):

l)absence of a stylistic option
2)substitution of an analytic construction for a synthetic one

3)analogical leveling
The difficulty with defining these changes as simplification that is characteristic of language decline is that they are fairly standard changes that occur in almost any language. Moreover, some of the changes like simplification of analyticity, for example, can be compensated by word order complications elsewhere in the language. (Hoenigswald 350)

The second kind of simplification, autonomous simplification, is even trickier, if it can be defined at all. A low rate of relativization IS one feature that most would consider to be unambiguously a simplification. However, in « The Social Functions of Reiativization », Hill points out that the declining rate of relativization in Mexicano is an aspect of solidarity coding and not of simplification as a result of language decline, as linguists would be prone to interpret. Solidarity coding is used in contexts in which the speakers already share a
great deal of information. Relativization, on the other hand, has a decontextualizing property and contains high information, which IS not characteristic of the kind/ of speech typically used among


speakers of low social distance. As use of Mexicano came to be restricted to contexts of intimacy, relativization decreased.

In this paper I am concerned with the causes of language death In Africa, particularly in East Africa. Language is a medium of social communication. The language that a person chooses to use in a specific context has much to do with hislher group identity. And group identities overlap in complicated ways in Africa. Language death can be explored by studying the interaction of group identities and personal langauge choice. When a speech community stops speaking a language, it usually involves a shift in group identity as well.

Multilingualism is the accepted Dorm in Africa. Three types of languages play a role. The first is the language of the former colonial powers. European languages, mainly English, French, and Portugese, have spread throughout all African countries and have become the official languages of many of those countries. However, the use of European languages have been mostly restricted to certain domains, such as higher education, politics and business and to a small number of people. Only ten percent or less of the rural African population have a considerable competence in any of these European languages. (Brezinger et aI, 19) The second type of language is the African lingua franca. Kiswahili is the most common African lingua franca, but there are a number of other such high-prestige, urban languages that an increasing number of speakers are speaking as a second language. The third type of language is the indigenous languages. Africa is characterized by an unusual richness and variation of indigenous languages, but these are in danger of being replaced, not


by the European languages, but by the African lingua franca. When discussing language death, it is almost certain to concern one of these indigenous languages. Brenzinger et al claim that there are 222 cases of languages/dialects in decline or on the brink of extinction in

Africa. Leaving aside the complicated question of where to draw the line between languages and dialects, the real number could be expected to be more, since the countries with reliable linguistic information tended to have a dis-proportionately high number.

Language and identity intersect at varying levels in Africa: national, ethnic, and personal levels. It is important to distinguish between these three levels of identities, because each denote very different approaches. Language is the common factor, because it IS what runs through all three levels. Karl Deutsch writes that « A community of language IS a community of information vehicles: most words of the language, or at least the words most frequently used, will be recognized and spoken by most members of the group, with identical or closely similar denotations » (41). He proposes a functional definition of nationality that we can easily adapt, for our purposes, to the multi-layered level. Ties of transportation, economic intercourse, social stratification, cultural similarity, and similarity in already existing speech habits, will all have their effects in determining what the actual speech community will be at any time. In a speech community, Deutsch sees « channels of culture », through which information, knowledge, values, and traditions are disseminated. The transmittance of information is important, because « a community consists of people who have learned to communicate with each other and to understand each other well beyond the mere interchange of goods and services » (91). These networks of communication set the community apart from others. In other words, the more vigorous the language, the denser the communication network, and hence a stronger sense of community among those who use the language. We can adapt this idea to see an overlapping of three levels of « speech communities » in Africa. One person may be a member of three levels, or less or more simultaneously, depending on how many indigenous languages she knows.


Talking about the nation-state in Africa is different from talking about the nation-state in Europe. The nation-state in Africa is actually only a state, in that it has administrative sovereignty over clearly defined boundaries. The government is not necessarily seen to be a legitimate representative of any « national sentiment » with roots in a shared tradition, values, and ancestral ties. Therefore, the nation-state is not a unit of identity that people m Africa usually identify with on a personal level; however, it does influence what language is used in the schools, and in government, which ultimately has an effect on people’s personal choice of languge.

The nation-state was an European invention that took root in Africa in the post-colonial era. It arose out of the arbitrary drawing of boundaries by the colonial powers. The precolonial political formations of Africa’s « tribal » past were actually communities with histories and traditions that were rooted in popular acceptance. They were identities to which people strongly held. For example, in pre-colonial Somali society, the ideology of kinship had two central elements: blood ties and Xeer. The first was essentially a product of genealogical connections, and the latter was the locus of a pan-Somali code of conduct. Blood ties aside, the Xeer was especially significant because it provided a social contract to check conflict and constrain centrifugal tendencies. (Samatar, 9) As the colonial powers moved In to set up administrative governments, tribalism came to take a different meaning. The new tribalism lost the notion of anything resembling the Xeer of Somalia; it was a convenient invention of the colonial period. Gathering Africans into invented tribes and favoring a few could make colonial administration easier and cheaper, while from an African standpoint, being so gathered could produce a stranger bargaining voice. (Davidson, 225) In other words, tribalism was the product of the « divide-and-rule » policies of the colonial powers to facilitate administrative rule. Also, the integration of the economy into the international economy generated a whole new class of merchants and an urban elite who fought for the appropriation of surplus generated by the pastoral producers in the countryside. The era of post-colonialism set off, in many African countries, a race for the spoils of government power as the colonizers were moving out. During this time, historically separate and distinct communities banded together and declared themselves to be new « tribes », and other new tribes were formed out of nowhere. Many were formed out of personal ambitions. These tribal associations later became rallying points for nation-state politics. The new African elites, who were a select group educated by the colonial system, came to think that the advancement toward the nation-state was the only escape from the colonial condition. This was due to the mentality of the elites, who thought of Africa’s history as being irrelevant. The decolonizing effort was an attempt by the elites to turn colonially formed territories into nation-statist territories. Taking control of the government offers huge rewards in terms of foreign aid and control of the whole government bureaucracy. The problem was that there had never been a center of power in the European sense in Africa within those arbitrarily drawn nation-state borders.

Language planning and imposition had been part of the colonial enterprise, even before nation-states were born in Africa. Among the preconditions for establishing colonial authority was communication with the colonized. Use of, and control over, language was needed to maintain economic, military, and religious regimes 10 power. At times it was in the interest of the colonizers not to establish a fully developed communicative network with the colonized, either by learning their language or by forcing them to learn a European language. This is seen in the way Swahili spread in the colonial eraas the lingua franca for colonial rule. Fabian says « It would seem that the decisive factor in the pidginization of Swahili documented in the context of industrial labor in Katanga was not so much a narrow range of functions or types of communicative interaction as the highly structured and hierarchical nature of interaction between European personnel and African workers » (l08). Pidgin Swahili may be a pseudo-form, resulting from relations of inequality and from active restriction of purpose/function. Use of Swahili by European supervisors in Katanga was tied to enactment of power in a hierarchical organization (i.e., giving orders and responding). By carefully rationing French for Africans and presenting Swahili to Europeans either as forbiddingly difficult or·as ridiculously easy, any free flow of exchanges that could have gone beyond the necessities of formalized relations was effectively discouraged.

More recently, language planning and policy have been part of Africa, as they have been in many developing countries. Once again one sees a relationship between language and power. Language planning and. policy is almost a necessity for the whole nation-state package, because nation-states require a central government with a bureaucratic apparatus. Even when the adoption of an official language is seen to be an administrative issue, language is a potent symbolic issue. By adopting an official language, the government is essentially claiming ownership to the education and bureaucratic apparatus. Depending on which language is chosen, it affirms the worth of the group claiming linkage with the language (Horowitz, 222) Policy choices have consequences, not just for careers of members of one or another ethnic group, but for social-class
mobility, for bureaucratic effectiveness, and for international contact. To opt for multilingualism usually means perpetuating the colonial language as the interethnic link, thereby preserving the advantages of the advanced group, with its greater mastery of that language, and also preserving ties to universities in the former metropole. To choose an indigenous language is to make job recruitment more egalitarian classwise, but to discriminate ethnically.

Efforts have been made to generalize the language issues facing devloping countries and to predict the outcome of a country’s choice of official language. There are three types of language decisions, according to Fishman. Type A decisions are those decisions that come about as a result of consenseus that there is neither an over- arching sociocultural past nor a usable political past that can currently serve integrative functions at the nationwide level. It is felt by elites that there is as yet no indigenous Great Tradition. They select a language (usually Western) of wider communication and the continuation of a Western trained elite. Type B deicisions are based upon long established socio-cultural unities and upon rather well established political boundaries as well. There is widespread consensus that a single Great Tradition is available to provide the

things for nationwide identification. They believe they already possess a strong national identity but must seek to render it more functional for the purposes of national well being in the modern world. Efforts are made to « modernize » a indigenous language. A choice must often be made between a highly sylized, classical variety of the language and a vernacular variety. Type C decisions are characterized by a multiplicity of Great Traditions. Their cooccurrence within a single policy makes for rather constant internal tensions and for nationalistic disunity, particularly in the absence of superordinate threat.

In the case of Africa, rarely do we find Type B decisions- and no wonder, considering the incredible complexity of languages in general. ‘added to the superficiality of the nation-state boundaries. Also, Fishman’s analysis only gives us two alternatives, No Great Tradition or Many Great Traditions, and predicts the unanimous adoption of a Western language for the first, and conflict for the second. This is not very useful. Sadly enough, not many people, especially the politicians, do not recogmze that there is something

equivalent to a Great Tradition in Africa, so they would probably dismiss the second alternative of Many Great Traditions.

Furthermore, even conceding the existence of Many Great Traditions in African countries, the outcome is not always conflict. The outcome seems to have to do more with the relative political strength of the politicians claiming to « represent » a ethnolinguistic group, rather than the existence or non-existence of Great Traditions. If there is one group that is numerically and relatively stronger than all other groups in the country, then that group’s language is likely to be adopted as the lingua franca. When groups within a country cannot agree on the lingua franca because their relative vitality is roughly equal, they will usually choose an European language for a lingua franca to avoid further conflict. Recently, moreover, there has been a renewed interest in promoting the use of African lingua francas, to avoid the use of European languages which are associated with the colonial era.

Three countries In East Africa have chosen different official languages, depending on the conditions facing each country.
Tanzania chose Swahili because Swahili was not associated with a numerically dominant or politically powerful group of native speakers. (Scotton, 203) There was a tradition of using Swahili as the main indigenous lingua franca across the entire country, and this tradition had bleached Swahili of much of its ethnicity, making it a neutral choice. Swahili appeared on the eastern coast of Africa about 1000 years ago as a Bantu language. There it came into contact with Arabic and Islam and remained a coastal language until the early nineteenth century.’ Until that time, trade proceeded along traditional routes, carrying goods from the interior of the continent, so Swahili remained a coastal language. Then, trading patterns changed and goods were carried from the coast to the interior. The new caravans carried cotton goods from the U.S. in exchange for ivory. (Wardhaugh, 192) At the same time, Christian missionaries set up operations in the coastal areas where Swahili was spoken. They contributed to the spread of Swahili as a lingua franca by codifying the language, and publishing grammar books.


In Uganda and Kenya, English has been adopted as the lingua franca. Much of the influence that English and French have in sub- Saharan Africa today derives from the ways in which the languages were promoted during the colonial era and from the roles they play in the world. The way that Engish was promoted in British colonies may have contributed to its popularity as a lingua franca. The colonial powers had different attitudes toward indigenous languages. The British colonizers were willing to learn the language. They also allowed native children to be educated in the local language (in the few schools that they had). There was no uniform policy; British education has always been heavily decentralized. One result was that former British colonies were left with two important attitudes toward language. One was that English was a pragmatic language, a language that should be learned because it was useful. The other was that vernacular languages themselves were of some worth, because education in indigenous languages often went on alongside education in English. This was in marked contrast to the French colonizers. French was the language of a high culture and an advanced civilization; all elites had to be educated in French. The French system tended to produce a very centralized education system for a very small number of educated elite who subscribed to the same values that the French did. From the above observation, it is possible to surmise that English was more « neutral » as a language than French and thus politically easier to implement as a official language.

The ethnic group is the second unit of analysis in studying the relationship between levels of identity and language. When an individual speaks a language the language serves both as a medium of social communication and as a outward sign of collective belonging to a group. One way to approach language death is to study the different pressures affecting ethnic groups. What happens to the ethnic group matters for the fate of the language, although to different degrees depending on the situation.

The definition of an ethnic group seems to depend on who one asks. In a broad sense, however, one can say that ethnic groups entail the existence of some kind of boundary separating the group from others, and some kind of defining culture that mayor may not be exclusive to the group. A more subtle analysis requires the use of both the objective and subjective definitions of ethnic identities.

The objectivist approach to the analysis of ethnicity says that ethnic boundaries can be drawn through the identification of discrete cultural institutions and processes. Chief among these tightly integrated cultural elements is the possession of a distinctive language that may serve as a daily language in use, or, alternatively, serve only as a language of ritual. Such a language provides an interactive locus for an ethnic group and allows for the


communication of umque group symbols. Language is the vehicle for a world view that makes the group different from all others. The objectivist approach considers language to be one of the primary defining characteristics of ethnic identity. This involuntary approach to group membership emphasizes common ties of socialization. The main difficulty with the approach is that it does not adequately explain the persistence of ethnicity across generations in different geographical areas and changing linguistic patterns, or within rapidly

changing social contexts. It is at this point that the subjective perspective is useful.

The subjectivist VIew claims that ethnicity reflects a shared feeling among a collectivity that may be internally differentiated. They say that ethnic boundaries are marked by lines of mutual recognition and reciprocity of exchange. This means that a group’s definition’ of itself is formed through interaction with other groups. Recognition and reciprocity is a differentiating process in relation to other groups. The subjectivist view argues that language and other seemingly objective variables are disposable symbols of an
underlying sense of peoplehood. « The politics of ethnicity may be fought around such objective issues as the language of instruction and administration, but the heart of the matter lies in the self- identification and political affirmation of a group of people who consider themselves to be different » (Ross 1979). But the sense of peoplehood need not be totally empty and subjective as an extreme view would seem to suggest. Ancestry is often assumed, even by the subjectivist view, to be a objective criteria for membership in the

ethnic group.


Some combination of both perspectives is necessary. We can say that ethnic identity is allegiance to a group, large or small, socially dominant or subordinate, with which one has ancestral links. There is no necessity for a continuation, over generations, of the same socialization or cultural patterns, but some sense of a group boundary must exist. This can be sustained by shared objective

characteristics or by more subjective contributions to a sense of groupness, or by some combination of both.

The vitality of an ethnic group is dependent upon status, demography, and institutional support. (Giles, Bourhis and Taylor, 309) These are variables that measure the degree to which an ethnic group can thrive as a group. In the writings of Giles, Bourhis and Taylor, there is an assumption that the ethnic group is also an ethno- linguistic group. Even if one cannot assume this in all cases, it is reasonable to say that if the ethnic group has a language, then the vitality of the language will in large part be detennined by the vitality of the ethnic group.

Status is the broadest category of the three variables. and has associated with it four types of status: economic, social,
sociohistorical, and language status. Economic status pertains to how much control the ethnic group has gained over its own economic activities. In a capitalist economy. the ethnic group associated with a urban area or way of living is more prestigious. Social status refers to the degree of esteem the ethnic group affords itself. Socio- historical status pertains to the collective history of the ethnic group.

It can be a symbolic mobilizing factor to bind the group together and provide a justification for group solidarity. Language status refers to


the perceived international status of the language used by the ethnic group.

The Demographic variables are those related to the sheer numbers of group members and their distribution throughout the territory. Those groups that remain in one traditional territorial area are often better able to retain their ethnic vitality than groups that are dispossessed of their land. The concentration of the members is also important, as dispersion of a group tends to weaken it. Also, numbers are important. This includes not only considerations of the absolute number of members, but the relative growth of the population as well.

Institutional Support variables refer to the extent to which an ethnic group receives formal and informal representation in the various institutions of a nation, region, or community. This encompasses such issues as whether an ethnic group is well represented at a state’s executive level, and whether an ethnic group has organized itself as a pressure group at a more informal level. Also, of crucial importance for the vitality of ethnic groups is the use of language in the state education and bureaucratic systems.

In the East African context, there are two main patterns of language death that can be analyzed at the ethnic group level. The first is language death as a result of inter-ethnic group contact. The context of inter-ethnic group contact determines whether or not language death occurs gradually (the « encroaching diglossia » pattern), or suddenly without an intermediate period of diglossia.
Dimmendaal gives an example of a typical case of gradual language death using the Kwegu, a hunter-gatherer community, whose


language was gradually replaced by that of the Mursi and Bodi, both pastoral peoples. The hunter-gatherer groups in eastern Africa often live in close contact with the pastoral or agricultural groups, with whom they have a symbiotic client-patron relationship. This occurs because hunter-gatherer bands tend to be small by necessity, and it is convenient to have a patron who can guarantee stability in times of difficulty. The Kwegu are in a socially inferior position relative to the Mursi and Bodi. The Kwegu speak their own language amongst themselves. They speak the language of the Mursi and Bodi in inter- ethnic situations, but the Mursi and Bodi do not speak the language of the Kwegu. After a period of diglossia where the Kwegu still

identified themselves as a hunter-gatherer group distinct from the Mursi and Bodi and spoke their own language amongst themselves, inter-ethnic marriages increased, and the Kwegu ceased to speak their language. (Dimmendaal, 18)

Cases of language death without an intermediate period of diglossia are also documented. Dimmendaal notes that these cases generally occur when the assimilating group sees an advantage in adopting the social values of the dominant group. In the case of the expansion of the Maa as a language of prestige, for example, the assimilating groups adopted the higher-prestige subsistence pattern (pastoralism) of the Maa. Maa customs and social values were also adopted, during a short period. The expansion of the Maa community is a probably a result of its overwhelming dominant culture.

The second pattern of language death in East Africa occurs as a result of contact between an ethnic group and a lingua franca. As a


result of economic changes, urban centers are becoming increasingly important as a source of employment. As urbanization proceeds, many people find themselves living in situations where they must communicate with people of other ethnic groups. A lingua franca is different from a dominant-group language, in that it is not associated with any specific ethnic-linguistic group. Such a language is more likely to be adopted by groups concerned about preserving their group identity. In many urban centers, the ethnic language is often maintained, its use being mostly restricted to the home domain and to communication among members of the same ethnic community. However, in cases where the ethnic group is weaker,

speakers may begin to switch over to the lingua franca in domains previously assigned to the ethnic language.

From looking at the two patterns of language death, one problematic presumption made in the past on linguistic/ethnic identity becomes clear: there is no isomorphic relationship between language on one hand, and ethnic identity and/or culture on the other. It is possible for an ethnic group to lose its traditional mother tongue without losing its sense of identity. Ethnic consciousness is not necessarily dependent upon the maintenance of a unique traditional language. Dimmendaal lists several cases where an assimilated group still retains its distinct identity, yet has lost its language altogether sometime in the past. In the larger group

containing many such smaller, assimilated groups, « the emergence of a socially united group with internal dialect or mode-of-subsistence differentiation, and a political system characterized by gerontocracy


…implies language shift and, to some extent, modification of earlier cultural norms » (27).

Examining the causes of language death at the ethnic-group level, however, have its limits. Ethnic-group level analysis does not capture the full complexity of an individual’s · language choices. The ethnic-group level analysis captured communication at the inter- group level. The speaker was speaking as a member of a particular ethnic group. But a speaker is also an individual, who makes language choices at the personal level. An analysis of interpersonal encounters is also necessary. We want to know how speakers « modulate both the content and form of their utterance to

communicate effectively and to express varying degrees of solidarity or distance with their interlocutors » (Bourhis, 117).

An individual speaker navigates among different linguistic codes depending on who she is talking to, in what context, and about what topic. This linguistic navigation is called code-switching. Code- switching is defined as the use of two or more linguistic varieties in the same conversation or interaction. The varieties may be anything from genetically unrelated languages to two styles of the same language.

Bourhis divides the determinants of code-switching into three factors: macro sociolinguistic factors, micro sociolinguistic factors, and social-psychological factors. Macro sociolinguistic factors refer to the intergroup relation at a level dissociated from that of the actual conversation: diglossia and bilingualism, ethnic group contact, linguistic territories, and language planning are all macro sociolinguistic factors. These factors have been discussed in the


above section. Micro sociolinguistic factors include the topic, setting, and purpose of conversation as well as characteristics of the interlocutors. Social psychological factors help explain why individual speakers use the speech strategies they do, in terms of moodes, motives, feelings, beliefs, and loyalties in ethnic interaction.

Micro sociolinguistic factors explain well the kind of code- switching that distinguishes in-group and out-group languages. The idea that a language system itself could have certain domains of use originated with Bernstein’s theoretical framework, in which he presented two ideal code types- restricted and elaborated. He observed that the middle class in England were socialized into the attitudes, knowledge, and skills which constituted the elaborated code. This code was sufficient for all levels of social activity and technical knowledge. For the working class, however, language was

not a medium of special significance as it was for the middle class. Bernstein thought that speech for the working class had the primary function of defining the nature of immediate role relationships. Communication took place at a lower, hence more « restricted » level compared to communication among the middle class. Moreover, Bernstein claimed that the restricted code governing this speech had a specifiable linguistic structure commensurate with its social functions. The grammar and lexicon of the restricted code lack the hierarchical and flexible structure of the elaborated code.

Although the conclusion of Bernstein’s analysis could be disputed, it does alert the linguist to the existence of multiple roles of language. One possibility is that the same language could take on different characteristics (codes) depending on whether it is used as a


in-group language or as a out-group language. Also, code-switching allows for differentiations that are not possible within the ethnic group-level analysis, because a speaker could switch codes even when speaking to a member of the same ethnic group. The notion of code switching puts ethnic group considerations alongside a whole slew of different factors that influence an individual’s choice of language.

There . are several micro sociolinguistic factors that affect code switching. One factor is the setting of the conversation. Non- standard accents and dialects are considered most appropriate in informal settings, whereas more standard speech styles are most appropriate in formal settings. Another is the topic of conversation. Research has shown that speakers tend to revert to their native accent, dialect, or language when they are discussing an exciting or

. stressful topic.
One of the most important determinants of code switching

behaviour in ethnic interaction is the linguistic competence of each speaker. The consequence of speakers being bilingual or bidialectal is that such speakers often generate social norms about the appropriateness of code switching which are different from those found among monolinguals. When bilingual or bidialectal speakers interact with obviously outgroup speakers they are commonly observed to switch from their local mother tongue speech style to that of the outgroup speaker. Amongst themselves in intragroup encounters, these bilingual or bidialectal speakers usually maintain the conversation in their shared mother tongue.


Code switching can also be determined by social psychological factors: These factors cause individuals to adapt their speech to each other on different linguistic levels. For example, in a relationship of inequality, the dominant-language speaker may adjust his/her
speech to the lower prestige language in order to gain acceptance; the opposite is also possible. Also, if a speaker is dissatisfied with his social identity and wishes to change his position, he may adopt the speech code associated with the social identity that he wants to be identified with. These processes involve a speech strategy, in which the individual speaker is actually adjusting her speech to the othe! speaker depending on how she decides she wants to be perceived. It

may involve either speech convergence or divergence.
Giles’ theory of speech accommodation is what explains the

motivation and social consequences which underlie changes in people’s speech codes. A basic postulate of the theory is that people are motivated to adjust their speech styles, or accommodate to others’ speech styles. Giles proposes that the extent to which individuals shift their speech style toward, or away from the speech style of their interlocutors is a mechanism by which social approval or disapproval is communicated. A shift in speech style toward that of another is termed convergence, whereas a shift away from the other’s style of speech represents divergence. Giles also notes that when two people meet there is a tendency for them to become more alike in their accents, speech rates, pause and utterance lengths and so forth. (Giles, 322)

According to Giles’ theory, therefore, social identity is very wrapped up with language, because a group’s evaluative attachment


to its membership is reflected in its feelings about its speech style. Language seems to be one attribute which is associated with enough complexity and range of options that it can be used as a primary marker of group membership and communication. Language is a central issue in intergroup situations because it can reflect the range of complexity of social categorization that people perceive. (Giles, 325)

Coming back to the topic of change and language death, the above analysis of code switching at the micro sociolinguistic and social psychological levels presents another concept for defining the ethnic/linguistic situation in East Africa. Multilingualism is the accepted norm, and not the exception, in Africa. The emerging pattern of language use indicates the common use of a colonial-era lingua franca together .with an indigenous lingua franca. Therefore, the causes of language death should be examined at the level of the individual speaker’s code-switching decisions. Two tentative

generalizations can be made. First, lingua franca languages that are detached from unambiguous associations with specific groups will tend to be more vigorous and widely used. This is obvious because such languages are easily accepted among everyone in a multilingual setting where each must learn a new language to communicate with each other. Second, the pattern of values and connotations emerging from code-switching behavior may provide a kind of template along the lines of which social groups may later become distinguished. Multilingualism may precede and so prepare the ground for a new kind of social differentiation.


In a study of Nairobi. Kenya, Parkin found that the use of
either English, Swahili and the vernaculars generates a corresponding set of stereotypes. When reasonably fluent English is spoken, it connotes relatively high educational level and socioeconomic status. Swahili has ambivalent connotations. It facilitates the expression of national identity. At the same time, the informal contexts in which people come to know Swahili and its secondary status to English can imply that those who use it in ethnically mixed situations cannot speak English and so lack education. Finally, the four ethnic vernculars (Kikuyu, Kamba, Luyia, and Luo) generally connote ethnic inclusiveness and solidarity when used among native speakers. The use of the vernaculars connotes feelings of exclusion and opposition when used in ethnically mixed contexts. (Parkin, 193)

Parkin also studied the behavior of older children 1D Nairobi. He observed that the playgroups of young children comprise all four ethnic groups of Luo, Luyia. Kikuyu and Kamba. Swahili is used in the playgroups, while the children speak their ethnic languages at home. Then, when they reach adolescence, the children gradually polarize into two groups, Luo with Luyia, and Kikuyu with Kamba. (197) Both groups use some Swahili, but the first group uses Luo more than Luyia among themselves, and the second group uses Kikuyu more than Kamba among themselves. What is interesting IS that each of the two groups further divide into two, between the « English speaking societies » and « Swahili speaking gangs ». The groups do not differ much in their actual language use or ability, or in their social activities; what differs is their expectations for the future. Those who can expect to continue their education join the


English speaking socities, and those who canot join the Swahili speaking gangs. The exaggerated nature of the English speaking society members’ claims to speak English better than Swahili, and the claims of the Swahili speaking gangs to speak Swahili better than English, suggests that what is really at issue is the emergence of differences regarding their prospects in the structure of urban opportunities. Personal identification determines the code-switching behavior of speakers.

Judging from this pattern, a likely development is the emergence of a predominantly English-speaking elite. Although they are likely to be largely recruited from the dominant Kikuyu ethnic group, the English-speaking group includes enough members of other ethnic groups to appear polyethnic. English is not regarded as the distinctive property of a particular group, nor does it represent the country’s colonial past. It is increasingly becoming the language of the urban elite.

One issue that has not yet been considered is the effect of bilingualism or code switching on the structure of either language. The declining language often changes its form before it disappears. It is common to find interference or convergence phenomena in the structure of a language in contact with others, but it is difficult to

judge whether these are signs of language shift or language death.
In « Language Convergence and Language Death as Social Processes », Woolard says that one cannot tell for sure that the social conditions, processes, and activities that affect a language’s form are the same as the social processes that encourage or discourage that language’s continued use. Most notably, she argues that there is no direct


relationship between language purity and language death or maintenance, at least where phonological conservatism is concerned. It is a common assumption to make that languages that are « strong » are better able to maintain their « purity », but this assumption IS apparently wrong. Woolard suggests that adapting languages are strong languages: the more contact interference seen, the more longevity and health we may expect. Linguistic conservatism is not a sufficient condition for, or predictor of, language maintenance. Situations where language loss is well advanced offer no insight into conditions for maintetnance, but they do show that purity and conservatism are no guarantee of endurance, since some of them are

dying out with all their phonological complexities intact. Those social situations which enforce language survival create demands that lead to innovation and adaptation of forms in a language.

There are two types of language interference in the ethnic languages that suggest that, even in the face of urbanization and encroachment of the lingua francas, at least for now, those languages are not dying. The languages are adapting in response to language contact, which suggests that those « social situations which enforce language survival » are at work. The first type of language interference and adaption is language interlarding as documented by Agheyisi. Language interlarding « is the special kind of mixed code which results from the interspersing of indigenous language speech with English elements » (97). It is a language phenomenon which has virtually become a stable part of the verbal repertoire of urban communities in Nigeria. Although Nigeria is not in East Africa, we can assume that similar language phenomena exist in multilingual


urban communities elsewhere in Africa. Unlike the code switching phenomenon in the linguistic behavior of bilinguals, in which the verbal material is perceived as constituting two separates codes, language interlarding (IS) is characterized by its usual perception as a single code. An example of IS in Edo is:

Dial enumber naa, n’ uinform ‘en Mr. Oseni ighe a approve encontracti nil ne. (Dial this number, and inform Mr. Oseni that we

approved the contract already) (Agheyisi, 100)
As seen from the above example, IS is a kind of speech where English vocabulary is dispersed on a Edo language. It assumes that the other person has a knowledge of both languages.

Multilingualism is a normal feature of urban communities. The repertoire of each town or city consists of two categories of
languages: those used for intragroup interaction, and those used esspecially for interaction between members of different ethnic groups. The second category comprises English and other languages used as lingua franca. English usually holds a special position of prestige. As more people learn English, there are varying norms of usage, ranging from Standard English to varieties bordering on pidgin English to types that are only locally intelligible. The numbers of middle class, educated urban dwellers who speak English are increasing. According to Agheyisi, this « has given rise to the

virtually undirectional process of diffusion of linguistic elements
from English into the other speech varieties represented in the urban verbal repertoire » (108).

The use of IS occurs m settings usually associated with the use of indigenous languages, where the speakers are comfortable with


each other’s ability to speak both the indigenous language and English. A wide range of IS styles is encountered. For example, it is often used when the topic is about a foreign or technical subject, and the indigenous vocabulary is not sufficient. IS is used more often by younger people. Many believe that IS constitutes a threat to the indigenous languages, but if we remember Woolard’s analysis that the key to language survival is language adaption, then perhaps we can be optimistic about the fate of the indigenous languages.

The second type of language interference and adaption IS English loan words in indigenous languages. According to one approach taken by Knappert (1970), loan words reflect the cultural history of the interaction between groups. When there is language contact, loan words travel from one language to the other. It is possible to recreate the route and the direction which a word or group or words has followed, by tracing their phonetic forms. If one knows the phonetic system for the languages concerned, it is generally possible to determine the direction of loaning. Knappert also takes the approach that loan words tend to be a group of words with meanings that describe a particular aspect of life of the giver

language’s culture. Loan words carry with it the culture of the giver language group. One can show the sphere of cultural influence .of the giver languages by mapping loan words. (Knapper, 87)

In a study of English loan words in Shona (spoken in Zimbabwe), Bernsten and Myers-Scotton say that loan words can expand the range of phonological and morphological features of the recipient language by accepting words from the giver languages without integration. Where there is integration, it is syntactically


relevant. (Bernsten and Myers-Scotton, 126) Moreover, they claim that additions to the lexicon may include not only loans for new cultural items but replacement words for core vocabulary as well. Cases such as these where there is borrowing of unassimilated words has been shown « to occur in later periods of language contact when there is a high degree of bilingualism in the speech community » because the bilinguals already control the features of the second language and they are able to bring in features that do not exist in the first language freely, unassimilated to the first language vernacular. Unassimilated borrowing from the second language to

the indigenous language is one feature of multilingual speech communities. Loan words have important implications for a theory of language, because it shows that language is not a highly integrated unitary system.

In conclusion, there are several generalizations that can be made about characteristics of language death in East Africa:

1) Decisions at the national level do not affect language death directly, but does indirectly influence the individual’s language pattern. The speaker speaks both the lingua franca and his/her indigenous language.

2) There are two patterns of language death at the ethnic group level. First, there is language death as the result of inter-ethnic group conflict. Second, there is language death as the result of an encroaching lingua franca on an indigenous language. A temporary period of diglossia often precedes language death.


3)At the personal level, code switching is observed. Language death may be unlikely where active adaption of the indigenous language IS taking place, as seen in IS-speech and unassimilated loan words.



Abdulaziz, M.H. « Tanzania’s national language policy and the rise of Swahili political culture. » Language Use and Social Change. Ed. W.H. Whiteley. London: Oxford University Press, 1971. 1160-178.

Agheyisi, Rebecca. « Language Interlarding in the Speech of Nigerians. » Language and Linguistic Problems in Africa. Ed. Paul F.A. Kotey and Haig Der-Houssikian. South Carolina: Hornbeam Press, Incorporated, 1976. 97-110.

Bernsten, Jan. « English and Shona in Zimbabwe. » World Englishes. 13.3: 411-418, December 1994.

Bernsten, Janice Graham and Carol Myers-Scotton. « English Loans in Shona: Consequences for Linguistic Systems. » International Journal of Sociology and Language 100/101 1993:125-148.

Bourhis, Richard Yvon. « Language in Ethnic Interaction: A Social Psychological Approach. » Language and Ethnic Relations. Ed. Giles and Saint-Jacques. New York: Pergamon Press, 1979. 117-142.

Bourhis, R.Y. and Giles. « The Language of Intergroup Distinctiveness. » Language, Ethnicity, and Intergroup Relations. Ed. Howard Giles.
New York: Academic Press, 1977. 119-134.

Brezinger, Matthias, Bernd Heine, and Gabriele Sommer. « Language Death in Africa. » Diogenes no153 Spring 1991: 19-41.

Criper, Clive and Peter Ladefoged. « Linguistic Complexity in Uganda. » Language Use and Social Change. Ed. W.H. Whiteley. London: Oxford University Press, 1971. 145-159.

Davidson, Basil. The Black Man’s Burden: Africa and the Curse of the Nation State. New York: Times Books, 1992.

Denison, Norman. « Language Death or Language Suicide? » International Journal of the Sociology of Languages 12: 13-22.

Deutsch, Karl Wolfgang. Nationalism and Social Communication. Cambridge, Massachusetts: M.I.T. Press, 1966.

Dimmendaal, Gerrit J. « On Language Death in Eastern Africa. »

Investigating Obsolescence: studies in language contraction and

death. Ed. Nancy Dorian. New York: Cambridge University Press, 1989. 13-32.

Dorian, Nancy C. « The Problem of Semi-Speaker in Language Death. » International Journal of the Sociol02Y of Lan2ua2e 12: 23-32.

Edwards, J.R. « Ethnic Identity and Bilingual Education. » Ethnicity. and Intergroup Relations. Ed. Howard Giles. New York: Academic Press, 1977. 253-279.

Edwards, John. Society and Identity. New York: Basil Blackwell Inc, 1985.

Fabian, Johannes. Language and Colonial Power: the appropriation of Swahili in the former Belgian Congo. 1880-1938. New York: Cambridge University Press, 1986.

Fishman, Joshua A. « Language and Ethnicity. » Ethnicity. and Intergroup Relations. Ed. Howard Giles. New York: Academic Press, 1977. 15-52.

Fishman, Joshua A. « National languages and languages of wider communication in the developing nations. » Use and Social Change. Ed. W.H. Whiteley. London: Oxford University Press, 1971. 27 -56.

Freund, Bill. The Making of Contemporary Africa; the development of African society since 1880. Bloomington: Indiana University Press, 1984.

Giles, H., R.Y. Bourhis and D.M. Taylor. « Towards a Theory of Language in Ethnic Group Relations. » Language. Ethnicity. and Relations. Ed. Howard Giles. New York: Academic Press, 1977. 307-325.

Greenberg, Joseph. Language. Culture. and Communication. Stanford, California: Stanford University Press, 1971.

Hoenigswald, Henry M. « Language obsolescence and language history: Matters of linearity, loss, and the like. » Investigating

Obsolescence: studies in language contraction and death. Ed. Nancy Dorian. New York: Cambridge University Press, 1989. 347-354.

Horowitz, Donald L. Ethnic Groups in Conflict. Berkeley: University of California Press, 1985.

Khlief, Bud B. « Insiders, Oursiders and Renegades: Towards a Classification of Ethnolinguistic Labels. » Language and Ethnic Relations. Ed. Giles and Saint-Jacques. New York: Pergamon Press, 1979. 159-172.

Knappert, Jan. « Contribution from the Study of Loanwords to the Cultural History of Africa. » and History in Africa. Ed. David Dalby. London: Frank Cass & Co, 1970.

Myers-Scotton, Carol. « Patterns of Bilingualism In East Africa (Uganda, Kenya, and Tanzania). » International Handbook of Bilingualism and Bilingual Education. Ed. Christina Paulston. Westport: Greenwood Press, 1988.

Nida, Eugene A. and William L. Wonderly. « Communication roles of languages in multilingual societies. » Language Use and Social Change. Ed. W.H. Whiteley. London: Oxford University Press, 1971. 57-74.

Parkin, D. « Emergent and Stabilized Multilingualism: Polyethnic Peer Groups in Urban Kenya. » Language. Ethnicity. and Intergroup Relations. Ed. Howard Giles. New York: Academic Press, 1977. 185- 209.

Ross, Jeffrey. « Language and the Mobilization of Ethnic Identity. » Language and Ethnic Relations. Ed. Giles and Saint-Jacques. New York: Pergamon Press, 1979. 1-14.

Rothchild, Donald and Victor A. Olorunsola, ed. State Versus Ethnic Claims: African Policy Dilemmas. Boulder, Colorado: Westview Press, 1983.

Ryan, E. Bouchard and M.A. Carranza. « Ingroup and Outgroup Reactions to Mexican American Language Varieties. » Language. Ethnicity, and Intergroup Relations. Ed. Howard Giles. New York: Academic Press, 1977. 59-80.

Samatar, Abdi Ismail. « Destruction of State and Society in Somalia: Beyond the Tribal Convention. » The Journal of Modern African Studies, 30, (1992): 625-641.

Silvey, Jonathan. « The Social Context of Language Usage. » Cultural Adaptation Within Modern Africa. Ed. S.H. Irvine and J.T. Sanders. New York: Teachers College Press, 1972.

Wardhaugh, Ronald. in Competition. New York: Basil Blackwell Inc, 1987.

Woolard, Kathryn A. « Language convergence and language death as social processes. » Obsolescence: studies in lan2ua/:e contraction and death. New York: Cambridge University Press, 1989. 355-368.

Wurm, Stephen A. « Language Death and Disappearance: Causes and Consequences. » no 153 Spring 1991: 1-17.