Sie sind auf Seite 1von 20

Linguistic Society of America

Neighbors and Lexical Borrowings Author(s): Carol Myers Scotton and John Okeju Reviewed work(s): Source: Language, Vol. 49, No. 4 (Dec., 1973), pp. 871-889 Published by: Linguistic Society of America Stable URL: http://www.jstor.org/stable/412066 . Accessed: 10/04/2012 13:29
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Linguistic Society of America is collaborating with JSTOR to digitize, preserve and extend access to Language.

http://www.jstor.org

NEIGHBORS AND LEXICAL BORROWINGS CAROL MYERSSCOTTON University of Nairobi and JOHNOKEJU Bishop Kitching College, Uganda

Even with the recent interest in correlating linguistic performance with sociological variables, the tendency in discussions of lexical borrowingis to focus on the words themselves and to ignore the process. This paper argues that the process of borrowing determines what is borrowed. Most borrowings may well be nouns for items new to the borrowing language. However, given sufficiently pervasive cultural contact, speakers may borrowcertain types of core vocabulary items in great numbers. Data to support these contentions come from a study of the dialects of Ateso, an East African language.1

Two unfortunate assumptions persist in most writings on lexical borrowings, but will be challenged here. The first assumption lies in the standard theoretical model for describing and explaining types of lexical borrowing, in that its center of interest is the end product, the borrowed words themselves. Haugen 1950 and Weinreich 1953 have already complained that the socio-cultural setting in which the borrowing takes place has received little attention in studies of borrowing. The model does, of course, assume that languages come into contact, and it is often noted in passing that the contact is between speakers rather than languages. Still, the model largely ignores the actual process of borrowing, what Diebold (1961:98) calls 'a sociological learning process, viz. bilingualism and acculturation.' A second standard assumption which we challenge is that lexical borrowings represent mainly new items to the culture of the borrowing language. Borrowings which infringe on the core vocabulary of the borrowing language seem to be rarely mentioned. Numerous studies of European languages have been carried out on these bases. Recent studies of other languages have also tended to follow the same lines, focusing on the results of language contact rather than the process, and mainly presenting new cultural items as examples of lexical borrowings. So traditional practice has continued, in spite of arguments that a new model is needed: e.g., Haugen (1950:271) writes, 'Talk of substrata and superstrata [within a language
analysis which led to this article. Without his help in checking the interview tapes, making notes about loans present, verifying all Ateso words used in the study, providing information on Standard Ateso as a reference, and finally in reading an earlier version of the manuscript, this study could not have been completed. Scotton, however, takes final responsibility for the content and interpretation. She also wishes to thank others who aided her in the field work: James Oporia-Ekwaro,her interviewer; Father G. Hoefnagels at the Catholic Church, Molo; Father J. H. Hilders of Soroti; Mr. Silver Obbo of Molo; Mr. and Mrs. Ivan Karp in Amukura; and Mr. Hiro Nagashima, who was doing anthropological research in Gweri in 1970. She also wishes to thank Einar Haugen and Edgar Polome for reading the manuscript and for their comments. The map was drawn by George S. Barkley. Field work was supported by the Ford Foundation-financed Survey of Language Use and Language Teaching in Eastern Africa. Some of the analysis was conducted under a grant from the American Association of University Women for 1970-71.

1The first author,CarolMyersScotton,wishesto thankJohn Okejufor his aid in the

871

872

LANGUAGE, VOLUME 49, NUMBER 4 (1973)

which has borrowed] must remain stratospheric unless we can found it solidly on the behavior of living observable speakers.' One result of the concentration on the borrowed items, rather than on the speakers, has been the idea that borrowings are more or less uniformly present across dialects. Fishman (1968, fn. 30) strongly attacks the end result of this:
As a result of linguistic disregard for this phenomenon of sub-sets of speakers whose verbal repertoires are situationally and functionally patterned ('structured'), the laundry-lists of examples of phonetic, grammatical, lexical and semantic 'interference' that have been published are socio-linguistically quite worthless and misleading. Their implication that most if not all members of the bilingual speech communities under study reveal the kinds and degree of 'interference' indicated by the examples listed is almost invariably wrong.

However, even given these strong and ably presented pleas, it is interesting that there have been very few articles published in linguistic journals in the past few years about lexical borrowing.2Even articles repeating the assumptions which we are questioning are rare; apparently the subject of borrowing is considered solved or even dead by many linguists. The data we present here, however, indicate that borrowing is not a closed book which warrants opening only for a few comments in an introductory linguistics course. The richness and complexity of the data available deserve serious analysis with a new orientation. Drawing data now from an African language, Ateso, we will argue that the standard model for borrowing is unbalanced, and that we must modify both it and our assumptions as to the type of items borrowed. Our main arguments are contained in the following hypotheses: (1) An adequate model for lexical borrowing must take into account the subgroups of speakers involved in the language-contact situation and their sociological profiles. Speakers of any language can be considered a single unit only in the sense that they share some common core of competence in that language. Otherwise, and especially in terms of performance, these speakers must be viewed as members of particular sub-groups within the larger community. Different social patterns characterize the sub-groups; different linguistic patterns characterize their linguistic performance. What this means in any discussion of borrowing is that we must differentiate more carefully between (a) borrowings into Language X as a whole, (b) borrowings which occur only in certain geographical dialects, and (c) borrowings in certain socio-economically based dialects which may cut across the lines of geographical dialects. While it is widely held that every language has its geographical and socio-economic dialects, the tendency persists to speak of borrowings into a language as if they were uniform. Our data show they are not. (2) While our data support the standard empirical observation that many borrowings represent items new to the culture of Language X (the lexical borrowing coming into the culture with the new item), we hypothesize that borrowings within the core vocabulary itself are also very common, given sufficiently exten2For a bibliography of articles, some of which deal with borrowing (although most focus on other sociolinguistic topics), see Fishman 1970. A bibliography of sociolinguisticallyoriented articles specifically on African languages is found in Heine 1970.One article which does bring in sociological origins in its discussion of loan words is Polomd 1968.

NEIGHBORS AND LEXICAL BORROWINGS

873

sive contact with another culture. Again, our data from certain Ateso dialects support this argument. Only certain types of core items are replaced, and these are, to be sure, items which are probably the most peripheral in the core. But our data show extensive borrowings into the core, including (a) free-form system morphemes,3 such as 'but', 'how much', 'even if'; (b) short phrases which can almost be viewed as single units, such as 'there is none', 'I don't know'; (c) dates and times; and also (d) some greetings. The essential point is that these borrowings always refer to concepts already present as active lexical items in the borrowing language. Many people, of course, have noted informally some such borrowings; the introduction of the Italian ciao 'goodbye' into the speech of younger speakers of many European languages is such an example. But our point is that borrowings of these core FLOATERS may occur with such frequency and in such quantity that this type of borrowing warrants the same systematic attention as that of words for new items.4 (3) Finally, a general hypothesis is proposed about the relation between a speaker's sub-group membership and his borrowing patterns. First, the borrowing of words for new items is most often done by members of two socio-economic sub-groups, which often overlap: (a) those of a relatively high educational level, and (b) those who have traveled most widely. In either case, it is obvious that these are the persons whose lives include the greatest opportunity to borrow, as it were, because of the wider horizons available to these groups. This general idea is something of a commonplace, but it connects with one of the hypotheses we wish to propose: that the borrowing of words for new cultural items depends heavily on an individual's micro-world. Second, a major type of borrowing, that of core vocabulary floaters, is likely to occur with ANY speaker-given the cultural contact-in a general dialect area, regardless of his socio-economic group. That is, we hypothesize that borrowing depends heavily on an individual's membership in a wider or macro-group (which may well be geographically de3 Bolinger 1968, for one, uses the term SYSTEM MORPHEME to describe those morphemes which signal relationships between words or phrases in a language. The free-form system morphemes are also sometimes called FUNCTION WORDS. 4 We call these words FLOATERS because they are not especially fixed in the structure of the donor language-i.e., they are usually words composed of a single morpheme (or, if more than one, there are usually no inflectional morphemes). Given their morphological structure, it seems easier to borrow them than it would be to borrow an inflected verb form, for example. Haugen, who has repeatedly emphasized the importance of the 'learning situation' in any assessment of loan words, notes that such free-form system morphemes were borrowed by Norwegian immigrants in the learning situation of entering a new life in America. He writes (1953:372if.), 'In English came to them [the Norwegian immigrants] all the excitement of a new and pulsing world which was being built around them ... In such situations they learned phrases that were not strictly necessary, but which contributed slangy bits of vividness to their speech, exclamations that brought back the gay informality of American social occasions. They learned easily to introduce their sentences with "Well-", to embellish them with "By golly" and "O.K." and to describe people and things with the standard cliches. of "cute" and "nice".'He also notes (1956:66) that'one of the resistances to borrowing is certainly contained in the degree of boundness or independence of the linguistic items. Exclamations, which are fully independent utterances, are among the most easily borrowed (i.e. learned) items.'

874

LANGUAGE,

VOLUME 49, NUMBER

4 (1973)

fined, as it mainly is in our data). In one case, borrowing is more an individual matter; i.e., individual lexical items are borrowed by specific types of individuals within a community. In the second case, borrowing is more a matter of general diffusion, i.e. the pervasive cultural contact between speakers of the borrowing socio-economic groups. Of our three language and the donor language(s), ACROSS hypotheses, this must surely be the most tentative. Yet our data certainly support the possible correlation of differences in types of borrowing with differences in magnitude of social contact.

Languages and dialects in Ateso-soeakina areas UGANDA


Ateso=

/
I J
I I

10#0ft 0

Bantu languages- Luganda, Lusoga, Lugwere,

Lunyole, Lugwe,Lumasaba, Lusamia, Luluyia

(Lukhayo, Lubukusu) Western Nilotic languagesEastern Nilotic languages-

[
Sebei

Kumam, Dhopadhola Ateso, Ngakarimojong,


FIGURE 1.

NEIGHBORS AND LEXICAL BORROWINGS

875

Our data to support these hypotheses come from Ateso, an Eastern Nilotic language,5 spoken by more than 800,000 persons in central and eastern Uganda and into western Kenya in East Africa; see the accompanying map. Over 90% of the Iteso live in Uganda,6 mostly in the governmental area called Teso District,7 of which Soroti is the administrative center.
1. WHY STUDY LOANS ATESO?Speakers of Ateso and closely related languages IN

make up less than 12% of the population of Uganda. But these Eastern Nilotic languages are different enough from other Ugandan languages that no other more widely spoken language can represent them; thus Ateso, as the Ugandan Eastern Nilotic language with the most speakers, is one of the six official vernacular languages for schools in Uganda.8 (English has been the only official language used in the conduct of government in Uganda since Independence.) The extent and type of borrowing into Ateso is of particular interest for several reasons. First, Ateso is spoken in several non-contiguous areas, thus all Ateso speakers do not form one geographical community. Second, in each area, the Iteso are something of a minority. If numerical and cultural position has any effect on the borrowing which a language does, the Iteso should be relatively heavy borrowers. The Iteso are numerically a small group and, unlike other small groups in the areas where they live, they have no numerous closely related neighbors to strengthen their position. Also, in recorded history, the Iteso have exercised no cultural dominance over any of their neighbors.9 Third, in all cases, the neighbors of the Iteso speak languages which are relatively or very distant from Ateso in a genetic sense. In the usual situation, at least some neighbors of most African languages groups speak a related language; however, except for the Ikarimojong in northeastern Uganda, the Iteso have no contiguous linguistic brothers. Elsewhere, they are surrounded either by Bantu speakers, who represent an entirely different language family, or in certain instances by both Bantu and Western Nilotic language speakers. The Western Nilotics are so remotely related to the Eastern Nilotic group that some linguists would contend any relationship is mainly a matter of borrowing, without genetic basis.10
5 We follow Greenberg1963in calling Ateso an Eastern Nilotic language. Languages most closely related to Ateso in Uganda are Ngakarimojong (about 200,000speakers), Sebei (about 60,000 speakers), and Kakwa (about 60,000 speakers). Related languages outside Uganda include Turkana, Toposa, the Nandi-Kipsigis group, and Masai. All these languages are sometimes referred to as Nilo-Hamitic. 6 Ateso is the name of the language; the people are called Iteso; a single male is an Etesot and a single female is an Atesot; and the area is called Teso. 7 There are about people in Teso District, of whom 480,000are Iteso and 60,000are 540,Q00 Kumam. This is the estimate of Father J. H. Hilders (personal communication, May 1970). 8 Ateso receives 12% hours per week of broadcasting time on Radio Uganda. It is the official medium of instruction in primary schools in Teso District, and literacy campaigns have been conducted by the government in Ateso. A government newspaper and religious literature are published in Ateso. (These data are from Ladefoged et al. 1971.) 9 See Lawrence 1957for a cultural history of the Iteso in Uganda. 10Investigators such as A. N. Tucker, M. Bryan, and G. W. B. Huntingford have argued that the languages in Greenberg's Eastern Nilotic group are 'mixed' languages with both Nilotic and Cushitic elements in their basic structures. Greenberg argues that these languages are basically Nilotic, but with Cushitic borrowings.

876

LANGUAGE, VOLUME 49, NUMBER 4 (1973)

2. DATAGATHERING. We have concentrated on two geographical dialects of Ateso which are most prone to borrowing. We do this because: (a) these two dialect groups are relatively removed geographically from the area where Standard Ateso is spoken; and (b) it can be argued they are as susceptible to cultural identification with the surrounding alien language groups as they are to identification with any GREATER ATESO culture. The specific dialect areas we studied were located in the general area of Tororo in eastern Uganda, and around the village of Amukura in South Teso District in Kenya's Western Province. Field work was conducted in May 1970 by Scotton. Twenty-two persons were interviewed in the Tororo area, and samples of their Ateso were recorded. Twenty-six persons around Amukura were interviewed. For comparative purposes, thirteen persons were also interviewed in central Uganda, where Standard Ateso is spoken and where the majority of Iteso live (Teso District). These speakers of Standard or near-Standard Ateso included some individuals from Ngora, the acknowledged home of Standard Ateso, and also some individuals living in Soroti, the largest city in the Ateso-speaking area. In each area, an attempt was made to stratify the sample in terms of age, education, and occupation. We mainly chose men as respondents, since African men have wider social contacts than African women. Persons were asked questions designed to elicit speech in Ateso about their everyday activities. Recorded interviews lasted an average of ten minutes.1' During the field work, we also tried to check on borrowings into Ateso by a more formal device. A basic list of 100 core vocabulary items was drawn up in English/Swahili; and individuals, or more often, groups of individuals (such as men gathered to drink beer) were asked to give the Ateso equivalents of these words. In several cases, we asked locally acknowledged experts on Ateso to complete the Ateso word list, or to give opinions about borrowings into Ateso in their own areas.'2 Our interest was only in loan words into Ateso, not in differences between the dialect areas in Ateso core vocabulary. 3. THE TORORO AREA. Cultural influences on the Ateso speakers around Tororo are especially complex, as described below. 3.1. In the rural areas, speakers of the minor Western Nilotic language, Dhopadhola, are very numerous. The Padhola, who all live in the Tororo area, make up about 1.6% of the population of Uganda, and probably number 160,000.13In many cases, Ateso speakers very decidedly live in a Padhola culture. Furthermore, many Iteso marry Padhola, more often than not adopting Dhopadhola as their home language.
11Interviewing was done by an Etesot from the Tororo area, James Oporia-Ekwaro, a Makerere University student at the time. Taped interviews were transcribed and checked for borrowings by Mr. Oporia-Ekwaro, and then by both authors. The second author, John Okeju, is himself an Etesot from the Soroti area and a speaker of Standard Ateso. 12 For example, Father J. H. Hilders, the joint author of the Standard Ateso grammar (Hilders & Lawrance 1957), kindly supplied us with Standard Ateso equivalents for the word list. 18Of 17 hours per week of broadcast time for the Western Nilotic languages in 1968, only half an hour was in Dhopadhola. Dhopadhola is not officially used in any schools, nor are there any government publications in it. There have been literacy campaigns in Dhopadhola, however.

AND LEXICALBORROWINGS NEIGHBORS

877

The general belief around Tororo is that the Etesot accepts Dhopadhola as his language (rather than his Padhola wife learning Ateso) because Ateso is too difficult for many adults to learn.l4 It is ultimately pointless to argue whether Ateso is in fact MORE DIFFICULT than any other language such as Dhopadhola, since we have no objective criteria to measure 'degree of complexity' across languages. But the point to note here is that both Iteso and non-Iteso in this area seem to BELIEVEAteso is more difficult than other languages.'5 Another reason why Tororo Iteso learn their neighbors' languages more readily than others learn Ateso may have to do with the Iteso's position in the community. Until very recently-and still today in some instances-the Iteso recognized that they were considered by non-Iteso somewhat as second-class citizens around Tororo and in western Kenya. Apparently, the more numerically dominant Iteso in the main Ateso dialect area never had to face this problem. But Iteso in the other areas, where they are more recent immigrants (within the last sixty years), have been looked down upon. Today, more Iteso around Tororo are gaining positions of importance in the political and economic community, but certainly Iteso had little influence in the past. Around Amukura in Kenya, the Iteso still believe their political representation in Kenya politics is less than they deserve-although, admittedly, many groups make this complaint. Still, the point is that Tororo Iteso readily state that their status is lower than that of their neighbors; the Iteso in both Tororo and Amukura seem to learn their neighbors' languages more frequently than vice versa, and this phenomenon is widely recognized.16 The Padhola influence is strong on all sides of Tororo, although especially to the immediate north in such villages as Molo and Kidoko. Here we found persons calling themselves Iteso, yet able to speak Ateso only with difficulty, so great was their acculturation as Dhopadhola speakers. In Molo, a group of Iteso men found drinking beer from communal pots were conversing in Dhopadhola. When asked for various vocabulary items from our word list in Ateso, they did switch to Ateso and could give Ateso equivalents. Several times, however, they conducted their discussion of the equivalents first in Dhopadhola; and sometimes they gave the Dhopadhola equivalent by mistake, apparently thinking it was Ateso. Of the 22 Iteso interviewed in the Tororo area, only three said they could speak no Dhopadhola at all, and several said they spoke Dhopadhola better than
In fact, Ateso is a difficult language, having a fair degree of inflection in its nominal and verbal systems, and including a system of gender (masculine, feminine, and neuter) with an accompanying system of agreement between nouns and their modifers or pro-forms. Tone is also a feature of Ateso, and is used both to differentiate lexical items and to signal grammatical relationships, such as differences between verbal tenses. 15 John Okeju points out that, in the Standard Ateso area around Soroti, OTHER ethnic groups are more likely to learn Ateso in a mixed situation than are the Iteso to learn their languages. E.g., members of the Bakenyi (a Bantu group) who are in close contact with Iteso are always the ones who bridge the linguistic gap by learning Ateso. Similarly, many of the Kumam (a West Nilotic group) learn Ateso. 16 It is certainly true that other minority linguistic groups have proved to be more exceptional second-language learners than their neighbors. This is the case of the Berber in northern Africa, who live in an Arabic-dominated society. Both this and the Iteso case would tend to support a hypothesis that facility in second-language learning depends heavily on psychological and practical considerations.
14

878

LANGUAGE, VOLUME 49, NUMBER 4 (1973)

Ateso. Many claimed that, of all the second languages they know, they were most fluent in Dhopadhola. 3.2. Even though the Padhola influence is strong, the Bantu presence in the Tororo area is probably the dominant one over time. This can be accounted for by a number of factors: (a) the sheer number of Bantu speakers in the area, although many of the individual language groups are small; (b) the pervading influence of Baganda culture (with its Luganda language) in all of Bantuspeaking Uganda; (c) a history of Luganda usage, originally under the British, as a language of administration, education, and religion, second only to English and often superseding English in the hinterlands. Despite the weakening of the Baganda political influence after the 1966 Revolution, Luganda continues as the official vernacular language in the Tororo-area schools, and is the medium of instruction for the first primary grades. In practice, some other languages, including Ateso, are used instead in some schools. Also, Luganda is probably the most generally used lingua franca, with all but one or two persons in our Tororo sample claiming to know at least some Luganda. Often, the question of whether Luganda influence on Tororo Ateso is direct or indirect is next to impossible to assess, and we do not attempt it here. This is so for two reasons: (1) Luganda also left its impression (usually first) on Standard Ateso, and many of the Luganda loans in Tororo Ateso were probably first borrowed into the Soroti-Ngora dialects, since there were Baganda in this area from 1898 on; (2) many words for new cultural items were borrowed into Luganda from Swahili. The numerous smaller groups of Bantu speakers who live as immediate neighbors of the Tororo Iteso also have influenced loans into Ateso. But because words in many of these languages, especially loan words, resemble those in Luganda, it is difficult to assess the directions of any influence or its extent. These groups include speakers of Lugwere (about 170,000), Lunyole (about 140,000), and Lusamia/Lugwe (about 130,000). North of Tororo, in the general area of Mbale, are speakers of yet another Bantu language, Lugisu (Lumasaba), numbering about 510,000.17And across the border in Kenya are speakers of the various Bantu languages which make up the Luluyia complex. Fewer of the respondents claimed to speak one of the minor Bantu languages in the area than claimed to speak Dhopadhola, Luganda, or Swahili. But still, most claimed either to speak or understand a little in at least one of these languages, and many claimed to speak several of them. One man had a Gwere mother, another a Gisu mother, another a Nyole mother, and still another was married to a Munyole. There seems to be less intermarriage with these Bantu groups than with the Padhola. Of the nine other Tororo-area men who were married, five had Teso wives and two had Padhola wives, with two having both Padhola and Teso wives. 3.3. Since it is the main lingua franca all over East Africa, Swahili exerts considerable influence on all other East African languages, including Ateso around
of these Bantu languages is an official vernacular in the schools. There have been literacy campaigns in Lugisu and in Lasamia/Lugwe, but not in the other languages. In 1968,41 hours per week of radio time on Radio Uganda went to Lugisu and 32 to Lusamia/ Lugwe, but none to Lugwere or Lunyole.
17 None

NEIGHBORS AND LEXICAL BORROWINGS

879

Tororo. Swahili, also a Bantu language, has been most influential as the medium through which words for new cultural items have entered other East African languages. Speakers of Swahili were among the first people to come into contact with foreigners in East Africa, with items and concepts new to the East African situation. These Swahili speakers were either coastal dwellers who spoke Swahili as a first language, or traders using Swahili in the up-country as a lingua franca. As these Swahili speakers came in contact with new items, first through the Arabs, then through adventurers such as Portuguese explorers, and finally through the colonizing and missionizing British, they often borrowed names of new items. Swahili is still actively borrowing many words from both Arabic and English, as well as occasional words from other languages. In turn, other East African languages borrow these words, sometimes through Swahili and occasionally directly from English. It seems the rule that East African languages prefer to borrow words for new items rather than make up words out of native morphemes. Few of our Tororo-area respondents could speak Swahili well; but all the men could speak it to some degree, and some with real fluency. Neither of the two women interviewed claimed to speak Swahili or would even try. Men who had traveled, or who worked in a situation where they often came into contact with men of other ethnic groups, usually could speak something approximating Standard Swahili, although still a rather distant approximation. Still, Swahili has no real official standing in Uganda. Its only official recognition comes from the fact that since 1971 it has been one of the many languages in which Radio Uganda broadcasts. It is used as a medium only in primary schools for the military, and is a subject in only a very few secondary schools. But English, the official language, is not known by large percentages of the population; thus Swahili, as a widely known language, is often used as a lingua franca.18 Thus, even though Swahili has only informal status in eastern Uganda, its function as a lingua franca means that Tororo Iteso find it useful to know; and it exerts an influence on the Ateso spoken there. 3.4. Finally, of course, we must consider the strong influence of English on languages in East Africa. Even though, in general, only the relatively well educated (with some secondary schooling) even CLAIM to know and use English, every person's life and language is affected by English and the culture of English speakers in the Tororo area. First, there is the matter of prestige. The speaking of English is associated with the higher ranges of the socio-economic scale. Second, there is the matter of the almost inevitable pervasiveness of Western culture as transmitted by the English language. Development and progress are measured in terms of European-based practices, for better or worse. With these practices come new terms: cement-making is described in terms of English loans; office and factory routine, and such new concepts as unions, are most naturally dis18 According to Ladefoged et al., 52%of the men surveyed in Uganda as a whole claimed to be able to hold a conversation in Swahili; 51% claimed the same ability in Luganda, but this figure includes the 16%of the population who speak Luganda as a first language. 28% claimed conversational ability in English. Scotton 1972 reports that 97% of the working population surveyed in Kampala claimed to know some Swahili. Many of these people were immigrants to Kampala from outlying areas.

880

VOLUME NUMBER4 (1973) LANGUAGE, 49,

cussed in the language of the culture which brought them, English. Even some men who are no more than peasant farmers become accustomed to using such words as hybridin referring to their corn crop. In some cases, even though a word for a similar process exists in the indigenous language, some speakers use an English loan for the foreign process, especially if the speaker knows English well. Thus, when talking about attending evening classes in accounting, a young Etesot, whom we interviewed in Ateso, used an English base taking and inflected it with Ateso afixes to produce e-taking-ai 'I was taking'. Although only five of the 22 people interviewed in the Tororo sample claimed to speak any English, the speech of all showed the effects of contact with English in the loan words present. 4. THE AMUKiUA AREA. the crow flies, Amukura Iteso in western Kenya As are less than 10 miles from the Iteso in Tororo town, but they are culturally much more distant than this. First, there is the matter of physical and political barriers, the Uganda-Kenya border and the Malaba River on the boundary. Also, unless he walks cross-country, the traveler faces up to a 50-mile road journey between Tororo and Amukura. During the rainy season, some roads-even the main ones-are impassable in this area. The Tororo Iteso and their Amukura brethren are further separated from each other by a band of Bantu peoples who form something of a unit, since they all speak dialects of a common written language, Luluyia. These Baluyia people, who live in Kenya's Western Province and spill over into Uganda, number just over one million, and are a relatively large language group by East African standards. The Baluyia who live in Uganda, southwest from Tororo and just over the border, are the Basamia. Immediately around Amukura live the Bakhayo. Another dialect group, the Babukusu, live northwest from Tororo in Kenya. In the midst of this Baluyia domain, about 28,000 Iteso live in the Kenya These Iteso of Amukura governmental area called South Teso, around Amukura.19 exist in a many-layered cultural milieu which is somewhat similar to that of the Tororo Iteso, but which is always different in degree and usually in kind. The following factors are present: 4.1. The Amukura Iteso themselves form the dominant culture in the immediate Amukura area: it is estimated that about 80% of the population in the Amukura sub-location are Iteso.20Around Tororo, the Iteso are only one of many groups and certainly not the most numerous one; but in Amukura, the Iteso domination is shown by their control of all the main local administrative posts such as chief, sub-chief, and agricultural officer. The weekly Amukura baraza (town meeting), with all participants Iteso, is held in Ateso; outsiders addressing the assembly speak Swahili.
19Somewhat to the north in Kenya, in Bungoma District, is an area called North Teso where about 15,000 Iteso live. We deal here only with South Teso, and specifically with Amukura sub-location. 20 Estimate by Ivan Karp, who was gathering data for an ethnological study of the Iteso in the area in 1968-70 (personal communication, May 1970).

NEIGHBORS AND LEXICAL BORROWINGS

881

4.2. As members of the larger Iteso community with its center in the SorotiNgora area, the Amukura Iteso are considerably more removed than their Tororo counterparts. First, they are geographically more distant from Tesoland. Second, as citizens of a different nation, they have little practical call to think of themselves as members of the greater Iteso community. Still, distinct cultural ties remain between the Amukura Iteso and Ugandan Iteso. The mere fact that Kenyan Iteso maintain the Ateso language and Ateso customs attests to this.21 Also, our data show that the Kenyan Iteso travel a good deal-especially the men-and often to areas of Uganda where Iteso live. Further, the Amukura Iteso reported they listen to Ateso broadcasts over Radio Uganda; several Amukura Iteso noted that some dialect divergences are narrowing and attributed this to listening to Standard Ateso on the radio. 4.3. Both the Amukura Iteso and the Tororo Iteso, once outside the home-or even in the home-live in a heterogeneous environment. For the Amukura Iteso, the milieu is somewhat less varied since almost all their neighbors are Baluyia, while around Tororo we find a number of groups, the Padhola and various small Bantu groups. Marriage of the Amukura Iteso with Luyia women is very common, perhaps even more common than inter-ethnic marriages around Tororo. This may be because the Bakhayo who live around Amukura are primarily one clan; and according to clan doctrine, a Mukhayo must marry outside the clan. Of the 14 men in our sample who reported having wives, six had Luyia wives. Of those having only one wife, seven were married to Teso women, but one was married to a Mukhayo. Six had more than one wife; of these, five had one Teso wife and also one or more Luyia wives; one had two Teso wives. Of the seven women interviewed, two were Luyia women (one a Mukabras and one a Mubukusu) married to Teso men; four were Iteso also married to Iteso; and one was unmarried. Around Tororo, we found that a mixed marriage, especially with a Padhola wife, invariably meant that the Etesot husband would learn his wife's language for use in the home. Such is not strictly the case around Amukura, perhaps because the larger and less fragmented nature of the Ateso community reinforces Ateso's position as the appropriate language. Still, most people in our Amukura sample said they spoke some Luluyia (usually Lukhayo), and in general the Iteso (men) still learned something of their Luyia wife's language.
21 The following evidence indicates that Amukura Iteso are indeed conscious of their ethnic identity, and that speaking their own language plays a part in preserving this identity. A discussion was taking place in English, in a bar near Amukura on a night in April 1970, about the whereabouts of a typewriter belonging to the local school. Then one person intervened and said in Ateso, 'When Iteso gather together as we are now, the commonest language used is English. Why is this so?' Another fairly important community figure, also an Etesot, added in Ateso, 'This is true, because I have heard most of our secrets being known by non-Iteso. The only way to make people of other tribes not know our secrets is by practicing to speak our own language when we are alone.' Then the first man went on, this time in Swahili, saying he did not see why some people should boast by using English. The word 'boast' offended some of the speakers, and a fight almost started. But people quickly finished their beer and left-one saying in parting, in Ateso, 'You are too silly to say that we boast by speaking English.' (Details of incident kindly supplied by Ivan Karp.)

882

LANGUAGE, VOLUME 49, NUMBER 4 (1973)

Even with this inter-marriage and accompanying knowledge of Luluyia dialects, the Amukura Iteso do not seem to be especially influenced by Luluyia in their production of Ateso. We noted that one man (with a Luyia wife) spoke his Ateso with an overlay of Lukhayo intonation patterns; but for the most part, the external influences on Amukura Ateso seem to come mostly from Swahili or English, which far outstrip Luluyia as the dominant cultural forces. Ateso spoken in Tororo is also most regularly influenced by the dominant Bantu language (in this case, Luganda) as well as English and Swahili, rather than by the culturally less prominent minor languages in the area. But we also noted that Dhopadhola exerts a great deal of influence on Tororo Ateso. Even so, loan words from Dhopadhola seem less frequent than from the other languages. Rather, the most distinctive influence of Dhopadhola is seen in code switching from Ateso to Dhopadhola for whole phrases or even sentences, while ostensibly speaking Ateso. 4.4. As members of a wider community, Amukura Iteso naturally think of themselves as Kenyans; and Swahili is the dominant carrier of GREATER KENYAN in CULTURE this rural area. English may be the official language of Kenya, but it is largely the vehicle of the relatively educated, who are not numerous in rural villages such as Amukura. Swahili, however, is the vehicle of all. Every respondent, except for one old woman with no schooling at all, professed to know some Swahili; and many showed that they spoke Swahili very well, several near to the Standard dialect. Local administrative officials (who also know English, we might note) spoke Swahili especially well, and said they used it constantly. The 1969 announcement that Swahili is to become an official Kenyan language in the near future may have spurred people to use more Swahili, and even to claim to know more Swahili than in the past. This was our general impression on travels in Kenya, and specifically in our field work around Amukura; other researchers have noted the same phenomenon.22 The Amukura Iteso, particularly those in administrative positions or other jobs in which they meet the wider public daily, contrast with the Tororo Iteso in their command of Swahili. Around Tororo, most Iteso claimed either Dhopadhola or Luganda as their 'best' second language; around Amukura, the claim was almost always Swahili. A sub-chief in the Tororo area, when asked about Swahili's future in Uganda, declared Swahili's potential as a national lingua franca; but he himself spoke very halting and sub-standard Swahili. 4.5. As in Tororo, the Amukura Iteso live in a cultural milieu dominated at certain levels by English. We found the same kind of English loans in both areas, almost always noun loans for new objects and concepts. More Iteso in Amukura claimed to know English than they did in Tororo, however. Of the 26 persons interviewed around Amukura, half claimed to speak English. All 13 (except for one man who had been a cook in European hotels in Kisumu) were educated to the Primary 8 level or above. We had fewer Iteso educated to this level in our Tororo sample.
22 Ivan Karp notes that his field assistant, who had previously claimed English as his 'best' second language, reversed himself and claimed Swahili as his second language after the KANU announcement about Swahili's future official status (personal communication,

1970).

NEIGHBORS AND LEXICAL BORROWINGS

883

5. TYPESOFBORROWING. Let us now turn our attention to specific examples of loan words found in the recorded Ateso of our informants. We can readily see how the difference in general cultural milieu, from the Tororo dialect area to the Amukura area and to the Standard dialect area, makes its presence felt in the incidence and types of loans which occur in each area. Further, although in this paper we put the most emphasis on these macro-group differences, we can also see how certain loan words are a feature only of certain sub-sets of Ateso speakers; e.g., we will see that the useof English verb loans is found only in educated Ateso. The first thing which strikes us about Ateso loans is a commonplace: relatively speaking, large numbers are noun loans, and the same ones can be found in all Ateso dialects studied, including Standard Ateso (Soroti-Kumi area). But the nature of these loans is such that the same ones, and with about the same incidence, would probably appear in many East African languages and even in many languages of the non-industrialized world. These are the nouns which are borrowed directly or indirectly, mainly from English, for new objects and concepts, such as e-soda 'soda pop', a-motoka'motor car', and (e)-batisimu 'baptism'. While we recognize the incidence of such loans and will discuss them, what is of more interest in our study is the type and incidence of loans into the core vocabulary of Ateso, borrowings for which operable words already exist in Ateso. Such borrowings support our hypothesis that, given strong enough cultural contact, groups of speakers of a language will make core borrowings as well as borrowings for words for new cultural items. Our data show the following patterns concerning borrowings into the core vocabulary: (1) There is relatively very little borrowing into the core vocabulary by respondents from the Standard Dialect area. In the transcriptions of interviews of seven persons from Soroti town, only two such borrowings occur: Swa. makosa 'problems' in place of the native Ateso a-tionus, and kede 'with' from Kumam instead of Ateso nepepe ka. Likewise, we found very few core borrowings in the speech of the six respondents interviewed in areas from Gweri to Kumi. Again we found kede; and paka (Swa. mpaka) 'until, up to' was used instead of Ateso aitodol. One respondent also used a-musiri (Lug. omusiri) 'garden plot' instead of the Ateso equivalent, amana. (2) This very limited borrowing into the core vocabulary contrasts sharply with that in the Tororo area and in the Kenya sample from Amukura. Further, we get contrast between these two areas themselves. But since both these dialect areas are geographically separated from the area where Standard Ateso is spoken, and since the cultural influences of neighbors are somewhat different, we had hypothesized that we would expect different patterns of borrowing. In general, this is what we found: of (a) In both areas, the VOLUME borrowing in general is greater than where Standard Ateso is spoken. That is, we find more noun loans, more verb loans, more alien self-standing system morphemes, and more code switching in referring to dates and times, and even in whole sentences. (b) The source for the majority of the loans varies in these two areas according to the dominant cultural influences. Around Tororo, most core loans come from Luganda and Dhopadhola; around Amukura, most core loans come from Swahili.

884

LANGUAGE, VOLUME 49, NUMBER 4 (1973)

(c) In both areas, noun and verb replacement loans (i.e. loans to take the place of words already existing in Ateso) are relatively infrequent, but occur with about the same incidence. The sources are the dominant cultural forces in the area; e.g., around Tororo we find a-mani (Lug. amani) 'strength' used instead of the Ateso agogong by one respondent. Also aki-jokite (Dhopadhola joko) occurs as in Kinyek akijokite ong 'Don't trouble me' instead of the Ateso equivalent, Kinyek aitapas or Kinyek aitican eong. Around Amukura, most such loans come from Swahili; thus we find e-bidi (Swa. bidii) 'effort, diligence' instead of the Ateso agogong. Alongside Ateso aki-tumor 'to be delayed', aki-celewar (Swa. -chelewa) is found.23Some replacements come from English, but relatively few: e-kapenta 'carpenter' instead of Ateso e-keon was recorded in Tororo. (d) We found very few loan translations in any of our samples. The main one found is 'thank you for working', based on Lug. weebale emirimu. In both the Amukura and Soroti areas, this occurs (with Ateso words) as eyalama aswam, although it is perhaps more common around Tororo, where yoga aswam is the Ateso equivalent. (e) Many more borrowings of the self-standing system morpheme type (core vocabulary floaters) occur in Amukura Ateso than in the Ateso spoken around Tororo. Almost all of those in Amukura are from Swahili; but those from Tororo come from Luganda, Dhopadhola, and Swahili. Some of the Amukura loans, such as bas (Swa. basi) 'it's finished, well', etc., also occur in other dialects of Ateso, including the Standard dialect. Others, such as saidi (Swa. zaidi) 'more' in place of Ateso noi, are very common around Amukura; but they would be virtually non-existent around Soroti, for example. Some examples of core vocabulary floaters which we did find around Tororo include the following. Naye 'but', from Luganda, is very common, as in this interesting sentence: Nesi da Etesot naye nesi Epalat 'He too is an Etesot, but he is a Padhola.' Sasa 'now' (Swahili) is widely used around Tororo and at least known around Soroti; we recorded this sentence at Molo north of Tororo: Sasa ekwaikini kwana eong ai? 'Now what can I do?' (The idea 'now' actually is expressed twice in this sentence, with the Swahili loan sasa and with the Ateso kwana.) We hypothesize that the large number of varied borrowings of this type are so much more prevalent in Amukura than in the Tororo area because of the undiluted and constant influence of Swahili in western Kenya. Further, the social norms which sanction the general use of Swahili give it a prestige second only to that of English. Around Tororo, no single second language would be in such general use as is Swahili around Amukura. Finally, since so many people understand Swahili there, the Ateso speaker who uses Swahili loans in his Ateso, consciously or unconsciously, is sure he will be understood. The Tororo Etesot using Dhopadhola loans, for example, may well be understood by other local Iteso, but the possibility of incomprehension is certainly greater than with the Amukura Etesot using Swahili.
28 In Ateso orthography, the voiceless alveo-palatal affricate is written as c; in Luganda and Swahili (as in English), it is written as ch.

NEIGHBORS AND LEXICAL BORROWINGS

885

Thus we hypothesize that there are fewer such loans in Tororo Ateso because there is no one dominating alien culture, as there is in Amukura, from which an Etesot can borrow such words and expect to be understood by all those to whom he speaks. Some such loans do creep into the Tororo Etesot's Ateso, of course; the young people especially like to show off their versatility. But, often as not, the Ateso equivalent seems favored over a loan. In both areas, although English has the highest actual prestige, extensive loans from English, especially core loans, must-in practicality-be restricted to a few sub-groups, since only educated Iteso would understand them. (f) Code switching for numerals and dates is common in both the Amukura and Tororo dialect areas. Around Tororo, the Ateso speaker sometimes switches to Luganda. Around Amukura, it is almost the rule that numbers and dates are spoken in Swahili. (g) Code switching for whole sentences is found only around Tororo, and only in the areas where the Padhola culture is very strong. Both Dhopadhola and Luganda sentences or phrases were recorded in this instance. We hypothesize that such extensive code switching is found around Tororo because it is here where the Ateso culture is most fragmented. The Tororo Iteso live in a pattern of interspersal with their alien neighbors, so that some Iteso have almost lost their Ateso identity. Around Amukura, and even more so in Teso District, of course, the Iteso tend to live more as one group, even though there are aliens not far away. But no real conclusions can be drawn here about code switching, since our data are limited. Code switching for as much as an entire sentence was found only in the speech of four persons of the 22 interviewed in the Tororo area. That is, when asked to speak Ateso in the formal interview situation, most respondents could make their Ateso a satisfactory vehicle of communication. But in a more informal situation, as in the beer-drinking session visited near Molo, almost everyone-all Iteso-switched constantly from Ateso to Dhopadhola, with Dhopadhola actually dominating. Further detailed study is obviously needed before specific conclusions can be suggested. In ON 6. DETAILS TYPESOF BORROWING. more detail, the types of borrowing which exist in our samples are as follows: 6.1. By far the greatest number of borrowings, grouping all dialects together, are nouns for objects or concepts new to the Ateso culture. Some of these are words of relatively long standing in Ateso (60 to 70 years perhaps), such as e-mudu 'gun' or e-didi 'religion' (from Lug. emmundu and eddiini). Others are much more recent and not yet assimilated into the Ateso phonotactic system, such as e-servingai 'service' (in a petrol station), from English. Some loans are very widespread; e.g., words such as e-paba/e-pamba 'cotton' or e-muceri 'rice' (from Swa. pamba and mchele or Lug. eppamba and omuceere)are commonplace in the speech of every rural respondent. Words associated with education or administration also occur frequently, such as e-somero 'school' in Standard and Tororo Ateso (from Lug. essomero), and sikuli (from English through Swahili) in Amukura Ateso; e-musolo 'tax' (from Lug. omusolo) in many dialects, even in

886

LANGUAGE, VOLUME 49, NUMBER 4 (1973)

Kenya, although it is e-solo in Standard Ateso; e-lukudo 'main road' (from Lug. olugudo) in Tororo Ateso, e-gudo in Standard Ateso and e-barabara (from Swa. barabara)or e-rotofor a lesser road in Amukura Ateso. Nouns for objects or concepts involving industrialization and schools almost always come from English, and sometimes are borrowed in a relatively non-incorporated form, e.g. ledger, e-staff, e-standard,formeka 'formica', and e-radio. Other very common loans deal with buying and selling, such as a-pesa/a-pesan 'money, monies' (from Swa. pesa or Lug. eppeesa). Some nouns occur only around Tororo, or conversely only around Amukura; e.g., e-gobola/e-gobolola'sub-county' (from Lug. eggombbolola)is found mainly only in the Tororo area. It used to be employed in Standard Ateso; but now an Ateso equivalent, e-tem, has become more popular, at least around Soroti. Also, katale/e-katale (Lug. ekatale) 'market' is mainly a Tororo borrowing. Around Amukura, e-sokoni (Swa. sokoni) is the common rendering. Also, barasa (Swa. baraza) 'meeting' is a common loan in Amukura Ateso only. 6.2. A few nouns are for objects or concepts which already exist in Ateso, but for which the existing Ateso words do not quite capture the modem meaning. E.g., e-biasara 'business' (Swa. biashara) is used instead of Ateso a-gwelanar, which has more the meaning of 'trade, bartering'. A-uto is used in Uganda Ateso for 'cooking oil' (Lug. wuto) instead of the Ateso a-rurwak, which means rather 'animal fat', formerly used for cooking. E-kooti 'court' has an Ateso counterpart e-dieket with the same general meaning; but e-kooti is equally common, and its use usually indicates a graver situation. A-saduku 'box' (from either Swa. sanduku or Luganda) is used today for a receptacle for small objects and clothes; the native Ateso for 'box' is e-tuja, but refers rather to a larger woven structure generally used as a granary. 6.3. A few loans are strictly replacements for nouns already existing in Ateso for the same object or idea. Rwothi 'chief' from Dhopadhola is common around Tororo instead of the Ateso e-jakait. Speakers from Amukura using e-sambai 'plot, farm' (Swa. shamba) could use Ateso a-mana 'garden plot' or e-manikor 'larger plot'. But Ateso a-mana is sometimes replaced (by a-musiri from Lug. omusiri in Uganda Ateso) because it may be mistaken for e-mana 'vagina'. E-transfer 'a transfer' is sometimes used instead of the Ateso ai-julio, especially by English-speaking Iteso. A-mucaalat 'lady' was long ago borrowed (from Lug. omukyala) as a term for the wife of an important person; but it is now used for any respectable woman, and thus often replaces a-beru. Maskini 'poor one' (Swahili) is widely used as a euphemism instead of the Ateso loican. A-banja 'debt' (Lug. ebanjja) was adopted around Tororo, first by the Padhola, whose word banja was then borrowed by the Iteso. It also occurs in Standard Ateso, but the native a-pesen is more common. E-musuja 'fever' (Lug. omusujja) is in common usage around Tororo instead of Ateso e-imidi; it is also used around Soroti by some Iteso. Mammaa 'mother' (perhaps from Dhopadhola or Luganda, but many languages have similar forms) is preferred by some Iteso around Tororo; cf. Ateso toto, still dominant in Standard Ateso. In the Molo area, a-codopingot 'a woman who has left her husband' is used by the younger genera-

NEIGHBORS AND LEXICAL BORROWINGS

887

tion (from Dhopadhola codo 'break' and pingo 'handcuffs'); a-kobois the native Ateso word preferred in Soroti. An Amukura respondent used e-gireid 'grade' for 'quality' where he might have used the Ateso equivalent a-doketait. Another respondent used e-musia 'end' (from Swa. mwisho), but Ateso e-sang is more common in Standard Ateso. Some replacements have been reversed, so that Ateso originals are now preferred. At least in the Soroti area, the Ateso e-muron 'doctor' is now taking over from the loans daktari (Swahili) and e-musawo (Lug. omusawo). Also, ecoloni 'latrine' (from Swa. chooni or Luganda) is in use in all areas today; but some consider it crude, and prefer the Ateso ai-pany. Thus we see that while English is ultimately the major donor language for new objects and concepts, borrowings which are replacements generally come from languages of neighbors rather than the more alien English. 6.4. Borrowed verbs are relatively few; in general, they stand for new concepts, although a few replace Ateso equivalents. No single language is the dominant donor. Representing a new concept is ai-tuuda/ai-tuud 'to sell' (from Lug. okutunda): although Ateso had the verb ai-swec, this has more a sense of bartering than of selling objects for money. Ai-puga (Lug. okwefuga) 'to rule' stands for a concept new to the Iteso, who had no central authority in traditional societies, but were organized around clans. Ai-siom 'to study' and ai-wadik 'to write' (both either from Swahili or Luganda) also stand for new concepts. Aki-filinga, as in akifilinga letters 'the filing of letters', also stands for a new idea. One replacement is ai-serving, which appears in the speech of one respondent in an inflected form, i-serving-aete 'you (pl.) serve (food)', where the Ateso i-jaanakin would have sufficed. 6.5. Finally, a large number of borrowings into the core vocabulary common to the Tororo and Amukura dialects fall into four groups: (a) Words for numbers or dates. One Tororo man, while speaking Ateso, gave his age in Luganda as asatu mu mwenda 'thirty-nine'. One respondent combined Swahili with Ateso in saying salasini ki-kang 'thirty-five' (Swa. thelathini 'thirty'), although the Ateso word for 'thirty', akaisauni, is quite possible. One Molo speaker referred to the years 1962 and 1963 in Luganda as nkaga mubiri, nkaga musatu. In and around Amukura we found that switching to Swahili for numerals and dates was almost the rule, regardless of socio-economic sub-group. Thus, at a baraza meeting of all Iteso conducted in Ateso, we heard such phrases as tarehe ishirini na tano 'the date of the 25th', shilingi moja 'one shilling', mwaka mwili 'two years' (Swa. miaka miwili), themanini 'ninety'. Lug. siminu 'fifty' was also used. Other Amukura respondents were just as likely to switch to Swahili for numerals. This practice of giving numerals in Swahili is so pervasive that one Amukura farmer, who found no need to use a single loan word in his interview, finally switched to Swahili to give his age as salasini na tisa 'thirty-nine'. The fact that English is so little used in this type of code switching is worth noting. As noted above, English tends to be reserved as a donor language for strictly new objects and concepts. English names are common for the months,

888

LANGUAGE, VOLUME 49, NUMBER 4 (1973)

however, and this is true in many East African languages. But this may be because many African languages, including Ateso, have very cumbersome ways of expressing the months. Even Swahili, which does employ a short form mwezi wa 'month of ...' as in mwezi wa tisa 'month of the ninth', frequently uses English loans such as Septemba. (b) Short self-contained phrases. Especially in the Amukura sample, we found numerous instances of short phrases from Swahili being used in Ateso, even though Ateso equivalents were readily available. Examples include Letepesa hapa 'Bring the money here', hakuna 'there is not', Si namna hii 'Isn't that so?' (c) Code switching for long phrases or whole sentences. Such switching was found in a marked degree only around Tororo, where several respondents switched to either Dhopadhola or Luganda. The extreme case was a man from the MoloKidoko area who was very weak in Ateso, and at times answered Ateso questions entirely in Dhopadhola and once in Luganda. Educated respondents from any dialect area occasionally produced English phrases, as in Ebo ajaatere siong ... e-strong one 'Yes, we do have a strong one' (referring to a trade union). (d) Self-standing system morphemes. The speech of Amukura respondents fairly bristles with such loans from Swahili, e.g. yaani 'that is to say', raisi (Swa. rahisi) 'easier', ata 'even, up to' (Swa. hata), alakini 'but' (Swa. lakini), tu 'just; no more'. A number of these same loans from Swahili, as well as some from other languages, were also found in the Tororo area, but not so frequently. Swahili examples in Tororo include bure 'uselessly' and tena 'again'. More frequently, as we would expect in the Tororo area, loans which are self-standing system morphemes come from either Luganda or Dhopadhola. E.g., while Swa. kabisa 'entirely' is very common in Ateso around Amukura, in the Tororo area a Luganda loan nyo, with the same meaning, is in wide usage. From Dhopadhola we find such examples as kada 'even if'. Dingding 'very much' (Nilotic origin?) is also common around Tororo. In eyalama dingding 'thank you very much' (combining Ateso and the loan), we find what seems a loan translation from English. We 7. CONCLUSION. have presented data from several dialects of Ateso that the following contentions: support (1) An adequate model for lexical borrowing must first consider the borrowing process itself, i.e. the nature and extent of cultural contacts, and how they vary according to sub-groups within the borrowing group. Only then does one examine the product of the borrowing, the words themselves. Previous studies of borrowing which have dealt with the process at all perhaps have been too preoccupied with the idea that the amount of borrowing which takes place is determined by the structures of the languages in contact.24 Although we do not address ourselves directly to this premise, our evidence from Ateso shows that the socio-cultural
4 Diebold (111) brings up the same point. He writes that, rather than assume that interference is determined by the structures of the two languages in contact, 'A more modest premise would merely state that the two language structures set up limits of variability to interference. My data strongly suggest that sociological factors like age of learning, the learning situation itself, and intergroup social relations are equally crucial factors in determining the form of Spanish-derived interference phenomena in Huave.'

NEIGHBORS AND LEXICAL BORROWINGS

889

context, not the structures involved, seems more important in the process. The languages from which Ateso has borrowed so heavily all have very alien structures. (2) Borrowings of certain types into the core vocabulary of a language need to be recognized, alongside borrowings for items new to the culture. That our research has revealed many such borrowings in certain Ateso dialects may reflect the fact that we studied tape recordings of actual linguistic performances. Ideally, of course, we should have made many more recordings in a variety of social situations. But at least we have tried to follow the suggestion of Labov 1970 and others who emphasize the social context of language. They have pointed out the need to get away from the old one-language-one-informant approach, and to base descriptive studies and theories about language on actual linguistic performance of a representative sample of speakers. (3) Finally (on the basis of data from only one language, to be sure), we hypothesize that even the type of items borrowed may well depend on the type of cultural contact. We suggest that lexical items for new cultural items tend to appear in the greatest number in the speech of individuals of specific socioeconomic groups, those who have education and/or have traveled widely. But we suggest that borrowings of the type of core items we have found, such as self-standing system morphemes ('but', 'always', etc.) are not restricted to the speech of any single sub-group of speakers within a dialect area, given sufficient cultural contact with speakers of the donor language. REFERENCES DWIGHT. 1968. Aspects of language.New York:Harcourt,Brace & World. BOLNGER, A. JR. DIEBOLD, RICHARD, 1961. Incipient bilingualism.Lg. 37.97-112. A. JOSHUA 1968. Sociolinguistic on LinFISHMAN, perspectives the study of bilingualism. guistics39.21-49. . 1970. Sociolinguistics. Rowley, Mass.: Newbury House. Linguistics29:1, Part II.

H. GREENBERG,JOSEPH 1963. The languages of Africa. International Journal of American

York. [Received19 May 1972.]

POLOM~, EDGAR. 1968. Lubumbashi Swahili. Journal of African Languages 7.14-25. CAROLMYERS.1972.Choosing a lingua franca in an African capital. Edmonton: SCOTTON, Linguistic Research. WEINREICH, URIEL. 1953. Languages in contact. New York: Linguistic Circle of New

. 1953.The Norwegian languagein America:a study in bilingualbehavior.PhiladelIndianaUniversity phia:Universityof PennsylvaniaPress. (Reissued,Bloomington: Press, 1969.) -- . 1956.Bilingualism the Americas: bibliography researchguide. in a and University, Ala.: AmericanDialect Society. 1970.Statusanduse of African Weltform HEINE,BERND. Miinchen: linguafrancas. Verlag. to HILDERS, J. H., and J. C. D. LAWRANCE.1957. An introduction the Ateso language. Nairobi: Eagle Press. 1970. The study of languagein its social context. Studium Generale LABOV, WILLIAM. 23.30-87. CRIPER. 1971.Language Uganda.Nairobi: in LADEFOGED, PETER; RUTH GLICK;and CLIVE OxfordUniversityPress. J. C. LAWRANCE, D. 1957.The Iteso. London:Oxford UniversityPress.

EINAR. 1950. Problemsof bilingualism.Lingua2.271-90. HAUGEN,

Das könnte Ihnen auch gefallen