The Kuliak Languages of Eastern Uganda


353 86 2MB

English Pages [91] Year 1976

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

The Kuliak Languages of Eastern Uganda

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

BERND HEINE

The Kuliak Languages of Eastern Uganda

Q

East African Publishing House SEMITISKA SEMINARIET I m

r tA

m.

*

*

F irst published in 1976 by the EA ST AFRICAN PUBLISHING HOUSE L usaka Close off Lusaka Road P.O . Box 30571, Nairobi, Kenya

© BERND HEINE, 1976

Set and printed Letterpress by Kenya Litho Ltd. P.O. Box 40775, Changainwe Road, Nairobi, Kenya.

Acknowledgemen ts The present study was made possible by a research grant from the German Research Foundation (Deutsche Forschungsgemeinschaft) and a generous grant from the University of Nairobi which enabled publication. A number of people have contributed in various ways to this work. In particular, I wish to express my thanks to the Govern­ ment of Uganda and the people of Karamoja District for their hospitality and co-operation. I am also deeply indebted to Professor Bethwell Alan Ogot, Director of the Institute of African Studies at the University of Nairobi, from whom I have always received support and encouragement in carrying the work to completion. Furthermore, I wish to thank Dr. Alan Jacobs and John Weatherby for their stimulating comments and help, and Mohamed Abdulaziz, Chairman of the Department of Linguistics and African Languages, who kindly released me from teaching during the period of my field research. Finally, I owe special thanks to my friend and colleague Dr. John Sharman who not only advised me on various professional problems, but who also never got tired of trying to improve my English. B.H, Nairobi June, 1972.

v

Contents Acknowledgements

V

Introduction

ix

1

THE TERM “KULIAK”

1

U

The K

u l ia k

L anguages

1.2

The R

esearch

1.3

T h e P r e s e n t St u d y

8

COMPARATIVE GRAMMAR

11

2.1

11 11 23 30 30 30 32

Sit u a t io n

so

fa r

P honology

2,11 Consonants 2.12 Vowels 2.13 Prosodology 2.131 Tone 2.132 Stress 2.14 The Proto-Kuliak System 2 .2

L e x ic o n

2.21 Comparative vocabulary 2.22 Non-regular correspondences 2.3

G

rammar

2.31 2.32 2.33 2.34

Order o f meaningful units The noun The verb Deictic categories

4

35 36 52 53 53 55 57 58

GENETIC CLASSIFICATION

59

3.1

L e x ic o s t a t t s t ic a l C o m p a r i s o n

59

3.2

S y s t e m a t ic C o m p a r is o n

61

3.21 The first split of Proto-Kuliak

61

3.22 Tepes and Nyang’i 3.23 Ik and “ Dorobo” 3.24 Summary

64 65 66

4.0

KULIAK LANGUAGES AND HISTORY

67

5.0

EXTERNAL RELATIONSHIPS

69

APPENDIX

73

REFERENCES

80

viii

Introduction For more than a century the question as to how the languages of Africa can be classified has attracted the attention of linguists, and numerous hypotheses have been put forward. And yet, the linguistic map of Africa still looks chaotic and none of the many classifications which have been devised so far has found general acceptance among Africanists. This applies even to Joseph Greenberg’s genetic classification (1963) which, although being the most important piece of work so far produced on this subject, offers a number of new insights but no real solutions. One of the main reasons for Greenberg’s failure to achieve the hoped-for goal is that his method of language comparison is not adequate to cope with the extraordinary diversity of linguistic forms found in Africa. At the same time, the work of Greenberg has revealed that any attempts at large-scale comparisons for the purpose of discovering genetic relationships are as yet premature: they are beyond the scope of the techniques and methods presently available to historical linguists. Another attempt at classifying African languages on a synchronic basis has been undertaken recently by David Dalby (1970). By applying a “practical method” of language comparison he proposes a referential classification which is based on a mixture of geo­ graphical and typological linguistic criteria. Dalby’s “referential classification” may prove useful to librarians, but hardly so to linguists. The author has failed to demonstrate in which way his “practical method” can contribute to our understanding of the linguistic relationships among the African languages1. As it stands, Dalby’s classification adds yet another to the many so far proposed and helps to increase the existing confusion rather than to resolve it. iThe linguistic criteria used by Dalby centre on such distinctions as those between gender and non-gender languages or between class and non-class languages which can be relevant to typological as well as to areal comparisons. The linguistic significance of these distinctions, however, is restricted and has already been overemphasized by several 19th century Africanists. IX

It seems that comparative linguistics in Africa today has reached a point where we must begin to think of its limitations. None of the comparative studies on a continental basis has proved successful. The alternative is to dispense with overall classifications and to concentrate on small-scale comparisons between closely related languages, both on the synchronic and on the diachronic level. Only if the relationships within smaller units have been established will it be possible to yield satisfactory results within larger groupings. The results obtained in this way are probably less spectacular but also are less likely to fall into early oblivion. At the same time certain beginnings in this direction have been made. David Dalby has himself presented a study of a group of closely related West African languages spoken in Guinea, Sierra Leone and Liberia (1965). This study is based on the application of the comparative method which up to now has remained the only reliable tool of historical linguistics for establishing genetic relationship between languages. Similar attempts have been made by authors of works on a couple of other West African language groups, e.g. J. M. Stewart with the Akan languages of Ghana, or B. Heine with the Togo remnant languages of Togo, Ghana, and Dahomey. By far the most successful work on these lines, however, has been done within the closely related group of Bantu languages. Authors such as Meinhof, Meeussen, Guthrie, and many others have raised the standard of research to a level which compares well with that of comparative research in other parts of the world, including that of Indo-EuTopean linguistics. Within East Africa, or more specifically, in the non-Bantu area of East Africa, research work of this type is still lacking. Instead of concentrating on small-scale comparisons between closely related languages such as Luo, Acholi, Lang’o, Alur, etc., most authors have rather preferred to engage in overall comparisons of groupings such as “Nilotic” or “Cushidc”, the linguistic relevance of which still remains to be demonstrated. In the following, a small group of East African languages sharing genetic relationship is considered in detail, with a view to investigating their history. It is hoped that similar studies on other non-Bantu language groups will follow.

X

1.0

T H E T E R M “ K U L IA K ”

The term “Kuliak” is used to refer to a group of languages spoken in the Karamoja District of Uganda. These languages are Tepes, Nyang’i, Ik, and a language called “Dorobo” by E. J. Wayland (1931)i. “ Kuliak” is the name given by lie-speaking groups to a population they met when they arrived in their present territory. Jie traditions relate that the Kuliak {ng'ikuliak ‘the poor people’) lived on or near a number of mountains and hills and were primarily hunters although they also were engaged in cultivation to some extent. It seems fairly certain that the Kuliak are the ancestors of most or all of the people inhabiting certain mountainous areas of Karamoja and speaking the languages mentioned above (see Lamphear, 1971). Although “Kuliak” is introduced here strictly as a term of historical linguistics, it may also possibly be used as a synchronic sociological term referring to certain linguistic minority groups which, due to similar ecological and historical conditions, share certain features of social behaviour12.

1.1

T h e K u l ia k L a n g u a g e s

The present study is based on a comparison of the following languages:

1This “Dorobo” is not to be confused with the “Dorobo” found in Kenya and Tanzania, 2The following remark made by C. M. Turnbull in his discussion of Ik society appears to be interesting in this connection: “The Ik regard all the neighbooring pastoral tribes with some contempt, but feel a kindred spirit to exist between themselves and the Napore and Niangea (i.e. Nyang’i; author’s note), to the west, and the Tepes to the south, referring to all, like themselves, as kwarikik, or “mountain people” .” (Turnbull, 1967:63)

1

(1) Tepes (Tepeth, spoken by the Tepes people who call them­ selves sir'at pi. sod. According to John Weatherby, the name “ Tepes” originated most probably among the Kalenjin peoples.1 T he Tepes occupy three extinct volcanic masses in southern Karamoja — Mount Kadam to the south, Mount Moxoto to the north, and Mount Napak to the west. According to a census made in 1964 they number about 2,000 (Weatherby, 1969:75), This figure is much higher than the one given by A. N. Tucker and M. A. Bryan, who estimate the number of Tepes to be “a few hundreds” only (1956: 93), but lower than the estimate of P.H. Gulliver who counted 3,100 on Moroto, 730 on Kadam, and some 100 people on Napak, giving a total of 3,930 (Gulliver, 1952:20). The Uganda Census of 1959 had given the number of Tepes people as 4,363 (Uganda Census, 1959:18). The Tepes language is threatened by extinction in the near future. On M ount Moroto it is spoken only by the older people while the younger generation use Karimojong as both their first and their primary language. On Kadam and Napak, however, this process of language shift is less advanced. (2) Nyang’i (Nyangiya, Nyangia, Nyangeya, Nyangea) spoken by a people who call themselves nyaifidt pi. nyatfi 2. They live on the southern ranges of the Nyangeya Mountains along the boundary of Karamoja and Acholi Districts. The Nyang’i speakers probably number hardly more than one hundred. The Nyang’i language is rapidly dying out. It is spoken only by people of over about 40 years of age. The population under 40 speak Dodos as their first language.3 (3) Ika (Ik, Teuzo, Teuth, Teuzo, Theuzo) spoken by a group of people who call themselves id’am pi. ika and their language icet'dda. The name teuso was given to them by the neighbouring Dodos, while the Turkana and Jie are said to call them ngulak (Tucker/Bryan, 1956:93). The Ik live in the northeastern edge

*The name given by the Kalenjin to Mt. Elgon, which at some time in history was inhabited by the Tepes, is tabasiat (Weatherby, 1969:124). 3According to Tucker and Bryan their own name is upale (1956:93). 3Already in 1932 J. H. Driberg reported: “There is no means of estimating the former size of the Nyangiya tribe, but it has dwindled now to a few hundred members only, who are rapidly forgetting their own language and prefer now to speak Dodoth . . . " (Driberg, 1932: 608).

2

of Karamoja District along the escarpment above Turkanaland, northwards from Kaabong, and over to the slopes of Mount Morungole, overlooking the northeastern corner of the Kidepo National Park (Turnbull, 1967:64). The number of speakers is not known. While Gulliver and Turnbull estimate it to be between 1,000 and 1,500 (Gulliver, 1952:20—approx. 1,150; 1953;97— 1,500; Turnbull, 1967:64—1,300), it "is calculated by missionaries to be not much more than 3,000" (Crazzolara, 1967:2). (4) "Dorobo”, spoken by the “Wanderobo”. The only inform­ ation we have on these people and their language comes from E. J. Wayland (1931). The language is probably extinct.

3

1.2 T h e R e s e a r c h S it u a t io n

so e a r

The first data on any of the Kuliak languages were published in 1931 by E. J. Wayland. On his third visit to Karamoja District he met a group of people whom he describes as mountain hunters and who called themselves Dorobo or Wanderobo (Wayland, 1931: 212-218). They lived in the mountainous country on the edge of the Karamoja-Turkana escarpment north of Kamion. Wayland was able to collect altogether 38 words of the Dorobo language. A comparison shows that there can be hardly any doubt that this language belongs to the Kuliak group, being especially closely related to Ik. The following possible cognates have been found: Dorobo (Wayland)

chiok hiss bugaten dack takaik emm gwa ho kesen km guass saba se watt ripp mes nakwitt

4

Ik (Heine) curuk ‘bull’

6is buk* ‘to marry* dakwa tak’aik ‘shoes* (Crazzolara) em gwah ho kzszn kua gwas ‘stone* saba ‘river* sea wat-on ‘to rain* r£ba may nyak uja

‘ox’ ‘spear* ‘marriage* ‘tree’ ‘sandal* ‘meat* ‘bird’ ‘house’ ‘shield’ ‘grass* ‘rock’ ‘valley’ ‘blood’ ‘rain* ‘millet’ ‘beer* ‘God*

In addition, there are three items which show correspondences, but which are probably loanwords from Karimojong. It seems, however, unlikely that Wayland’s Dorobo was a language identical with present-day spoken Ik. Of the 34 items having the same meanings only 13 are cognates. Even if one takes into account that the language has changed since then, and that Wayland was not always able to get the correct equivalent, it appears that we are dealing with two different languages or two very divergent dialects o f one language1. Shortly after Wayland’s publication, data on another Kuliak language became available. In his article “Lotuxo Dialects” J. H. Driberg published a list of 106 words of Nyang’i, called Nyangiya by him (1932:604-6Q5), The language used by Driberg in communicat­ ing with the Nyang’i people was Didinga. It seems that the com­ munication was not perfect as a number of misunderstandings can be observed. Driberg notes, for example, that the nominal plural is invariably formed by a prefix napat, which however is an adjective meaning ‘many’ (1932:609), while the plural o f nouns is formed in Nyang’i by a large number of suffixes. Driberg expressed the view that the Nyang’i might be related culturally and linguistically to a number of peoples inhabiting the mountainous areas further north, such as the Dongotono, Lokadhan, Lopit, and Lokoya, but he failed to give any concrete evidence for this2. In 1949 and 1950 Joseph Greenberg published his first version of the classification of African languages in a series of articles appearing in the Southwestern Journal o f Anthropology, One of the 16 African language families which he distinguished was called Nyangiya and consisted of Nyang’i as its only member. He remarked: “Among the dialects of .the Latuka of East Africa recorded by

1Cf. C. M. Turnbull’s statement: “The vocabulary draws widely from the languages of all the neighbouring tribes, and many of the words in Wayland’s word list are in fact not Ik at all.” (Turnbull. 1967:63) 2“The vocabularies suggest the faint possibility that in the distant past all these mountain tribes were related and spoke a common language, but that they have been isolated from each other by later Lotuko-speakm^ immigrants, and as a result of this isolation have largely fallen under the linguistic influence of their neighbours. There is not sufficient evidence to say whether their original language (if one may be predicted) was Hamitic or not, but in view of the absence of grammatical gender in Nyangiya (apparently the most primitive) the presumption is that it was not.” (Driberg, 1932: 603)

5

Driberg is one, Nyangiya, which, as Driberg himself saw, cannot be considered as Latuka or, indeed, as forming a part of any other language family of Africa.” (Greenberg, 1950; 1955:98). It seems that Greenberg did not have access to Wayland’s notes on the Dorobo language of the Karamoja-Turkana escarpment. Four years later Greenberg revised his classification and reduced the number of distinct language families to 12. Nyangiya was, however, retained as a separate family (Greenberg, 1954; 1955: 110). A short note on the Ik (Teuso) and Tepes was published in 1952 by P. H. Gulliver in an article entitled “The Karamojong Cluster” . Gulliver was able to collect a vocabulary of Ik and stated that this vocabulary (which apparently he did not publish) did “not agree with that of any other East African language . . . ” Concerning Tepes he remarked that this language is unintelligible to both Karimojong and Pokot (Suk) speakers (Gulliver, 1952:20). Another attempt at classifying the Kuliak languages was under­ taken in 1956 by A. N. Tucker and M. A. Bryan in their handbook volume "The Non-Bantu Languages of North-Eastern Africa” (Tucker/Bryan, 1956). The authors based their conclusions not only on the evidence published by Wayland and Driberg, but were also able to consult unpublished materials such as Dakin’s “Vocabulary of about 60 words of Teuso, Tepes, Nyangeya” (Dakin, s.a.) and King’s “Short vocabulary of Teuso” (King, s.a.). Tucker and Bryan established an isolated language group which they called “Teuso” and which consisted of the following languages: 1. Ik (Teuso) 2. “Dorobo” (based on Wayland) 3. Tepes 4. Nyang’i (Nyangeya) (Tucker/Bryan, 1956:93) Thus, it seems that Tucker and Bryan were the first ones to recognize the relationship between the Kuliak languages, although they gave no linguistic evidence to support their hypothesis. Concerning the external relationship of the Kuliak languages, Tucker and Bryan came to the same conclusion as Greenberg, i.e. that these languages are “unrelated to any other” {op. cit.). Ten years later, in 1966, Tucker and Bryan published a short grammatical sketch of Tepes (1966: 392-401) which was based on the evaluation of a questionnaire filled in “by a school-boy who did not really know his own language, but took the questionnaire

6

home to his village to fill in”. In addition, vocabulary material, partly recorded on tape, was available to the authors. In spite of the insufficient data, Tucker and Bryan succeeded in working out the fundamentals of Tepes structure. No information however is given on features such as tone or stress. In July, 1967, Fr. J. P. Crazzolara completed a manuscript on Ik after one month’s field research. The manuscript contains a 30page grammatical description and a vocabulary of approximately 1,000 entries. The phonetic and tonetic notation is excellent. Crazzolara’s work constitutes the most important work written on any of the Karamoja remnant languages. Greenberg’s classification of the African languages of 1963 differs from his earlier classification with respect to the Kuliak languages in two instances. As well as Nyang’i, Ik is listed. While in 1950 and 1954 he could not discover a genetic relationship to any other language group in Africa, here he included Nyang’i and I k a within the Eastern Sudanic sub-family, which belongs to his ChariNile branch of the Nilo-Saharan family (Greenberg, 1963:85-86). Greenberg justifies this step by including ten words taken from Nyang’i1) in his Eastern Sudanic Comparative Word List (op. cit. 95-108), By no means, however, can he claim to have proved a genetic relationship, as the number of items compared is far from sufficient and hardly any of his correspondences are phonetically convinc­ ing2. An alternative classification to that of Greenberg (1963) was proposed by A. N. Tucker (1967; 1967a). He suggested that the allocation of Nyang’i and Ik by Greenberg to his Eastern Sudanic sub-family “is perhaps attributable to the large number of Karimojong borrowings in Ik a, and Pakoot borrowings in Tepes.” (Tucker, 1967:679). Tucker came to the conclusion that Ik “could with advantage be reinvestigated for possible Erythraic affinities” (1967:660); and he classified it, together with a number o f other languages (which Greenberg had allocated to the Cushitic branch of Afroasiatic or to the Khoisan family), as “Fringe Cushitic” (see i.a. Tucker, 1967:661.) Tucker noted certain features which lOne of these words, o pi. oik ‘house’ seems to be an Ik word (ho pi. ho'ik) rather than a Nyang’i word (ir pi. irkdri). BCf. B. Heine, 1971: “The Eastern Sudanic Sub-Family—A Linguistic Relation­ ship?” Seminar Paper, “East Africa and the Nile Valley” , Department of History, University of Nairobi.

7

Ik has in common with Ancient Egyptian and Berber, but not with what he calls “Orthodox Cushitic” (1967:664; 1967a:21). In his comparisons Tucker relies heavily on Ik, but also uses linguistic evidence from Tepes which, according to him, “would seem to be grammatically a broken-down form o f Ik, with the loss of many significant elements” (1967:678). Nyang’i is not treated by him, probably because of lack o f data. In addition, Tucker discovered lexical affinities between Ik and Iraqw1, a language classified as Southern Cushitic by Greenberg and as “Fringe Cushitic” by Tucker. Tucker’s attempts at classifying the Kuliak languages are not more convincing than those of Greenberg. Although he had more linguistic data at his disposal, his method of comparison does not seem appropriate to prove a genetic relationship. The criteria used are largely confined to the system of free and bound personal pronouns and amount to a mixture of typological considerations and assump­ tions on sound-meaning resemblances. During the last five years, a number of articles have appeared on various Kuliak languages, thus extending our knowledge of the grammatical and lexical structure of these languages consid­ erably. A detailed analysis of Ik grammar has been published by A.N. Tucker (1971, 1972, 1973), supplemented in 1976 by a histo­ rical account and a wordlist of the language (Heine 1976). Gram­ matical and lexical notes on Tepes and Nyang’i appeared in the previous year (Heine 1975).

1.3 T h e P r e s e n t S t u d y The preceding discussion has made apparent the unsatisfactory state of research on the Kuliak languages. So far none of these languages has been described adequately—no grammar or dictionary of any of them has been published.

1“In vocabulary, however, the two languages have some significant corres­ pondences which, owing to the great distance separating the tribes, could not be discounted as borrowings: e.g, cow = lo in. Iraqw, /$> in Ika, //J in Tepes.” (Tucker, 1967:679)

8

This applies equally to the synchronic and the comparative diachronic aspect. That the Kuliak languages are genetically related to each other has been claimed by Tucker and Bryan, but linguistic evidence is still missing. The question of the historical relationship between Kuliak and other African languages and language groups has been the subject of a controversy between Tucker and Greenberg who both base their hypotheses on scanty, insufficient data. The present study aims at investigating the historical inter­ relationship between the Kuliak languages by applying the compara­ tive method of diachronic linguistics. Furthermore, it is intended to look into the question of how to relate the Kuliak languages to other language groupings in Eastern Africa. The linguistic data presented were, unless otherwise stated, collected during three research trips to Karamoja District between August 1970, and April 1971. During these trips, a grammatical sketch plus a vocabulary of approximately one thousand entries were collected for each of the languages concerned. Altogether about twenty linguistic informants were used for the research on the Kuliak languages. The main informants were:

Tepes (abbreviated Tp) Mariko Auca from Akeme (a village near Moroto Town), 26 years old, gardener. Nycmg'i (abbreviated Ny) Moding Liyamoe Luchul Lazaro Lokapil Narucha

all three from Lopalang’it, between 40 and 50 years old, farmers.

Ik Thomas Peter Tukei from Kamion, 24 years old, sub-county clerk. ” Dorobo” (abbreviated Do) no informants available. The orthography in this study is phonemic unless otherwise stated. Linguistic data from other than own sources are written in the orthographies used by the respective authors. Thus, for example,

9

words followed by Way (standing for Wayland), Dri (Driberg), or Cra (Crazzolara) have been taken directly from these authors. The orthography used is, unless otherwise indicated, based on that proposed by the International Phonetic Association.

10

2.0 COMPARATIVE GRAMMAR In the following chapter, the Kuliak languages will be compared with a view to reconstruct a common hypothetical ancestor language, called “Proto-Kuliak” (abbr. PK). The comparisons are divided into three sections. The first section (2.1) attempts to establish sets of regular phoneme-correspondences between the different languages in order to. reconstruct the phono­ logical system of Proto-Kuliak. Based on these correspondences is the reconstruction of morphemes and words, which is the concern of the second section.(2.2). The third section (2,3) then tries to investigate the rules governing the combination of words and mor­ phemes. The main aim here is to reconstruct features of the syntactic surface structure ofProto-Kuliak.Thecomparativemethod, on which this work has to rely, has not yet been developed to a point where accurate syntactic reconstructions comparable to those made in phonology are possible; hence, the results of the third section have to be considered somewhat more tentative than those of the preced­ ing sections. An analysis of the deep structure of the Kuliak langu­ ages is still lacking. The comparison of verbs is, unless otherwise stated, based on the imperative stem (2nd person singular) of the verb. 2.1 2,11

P h o n o lo g y

Consonants

*161 There is a series of voiced implosive bilabial stops for which a proto-phoneme */6/ is set up. The correspondences are Ik }6j: T p /6/ : N y m Ik 6etsr*: Tp 6sj: Ny b&j ‘white’ Ik ha6: Tp a6: Ny ab ‘hot’ Phonetically, the Nyang’i equivalent is also an implosive [6] which is an allophone of /b/.

11

Ik 6is: Ny bis ‘spear* Ik bar: Tp bar ‘sour’ W A proto-phoneme *jbj is reconstructed for a series of voiced plosives which are Ik /b/: Tp /b/: Ny jbj. As the phoneme jbj of Tepes has a free variant which is voiceless, the equivalent of this language can be either [b] or [pj. The same is found In a few examples of Nyang’i. Ik ber: Tp b&r: Ny bzr ‘to build, to mould* Ik boss Tp bes: Ny bes ‘ear’ Ik b\6sdn: Ny bos\dn ‘blue’ Tp rib: Ny rib ‘sky’

*/c/ The voiceless palatal stops jcf of Tepes and Nyang’i correspond regularly to the alveolar affricate /ts/ of Ik. For this series a ProtoKuliak phoneme */c/ is reconstructed, Ik ats: Tp dc: Ny dc ‘to come’ Ik tsak'fid: Tp cegd-dt: Ny cekud ‘stick for drilling fire’ Ik bits\in: Tp bed in: Ny bec\ in ‘leftside’

*/c7 Ik has an alveolar ejective /ts’/ which corresponds regularly to a voiced implosive stop j belonging to the phoneme j]j in Tepes and Nyang’i. A structural analysis of the phonological system of the Kuliak languages suggests that this series is derived from a voiceless palatal ejective */c’/ of Proto-Kuliak (see 2.14; 3.21). Ik ts'oot: Tp juj\ut: Ny j\ ujut ‘to suck* Ik bets' a: Tp 6&J: Ny bsj ‘white’ Ik silts**: Tp sfj ‘hair* Ik ts'an: Ny jan\at ‘louse’ The developments of Ik

can be described by the following rule involving the change of one distinctive feature: diffuse * — voiced _4- strident

12

>

[-f- diffuse]

The series Ik /ts */: Tp /j/: Ny /j/ does, however, not occur morpheme-internally between vowels. In this environment another series Ik /s /: Tp /s /: Ny /s/ is found. Ik kzsen: Tp kzstn ‘shield’ Ik noosan: Ny nausan\zt ‘clever* Ik b\6san: Ny bos\an ‘blue* The phonological shape would suggest that this series is derived from Proto-Kuliak */s/ (see below). However, as will be shown, */s/ has become /r/ in Ik intervocally. According to the most plausible hypothesis, the series Ik /s/: Tp /s/: Ny Jsj is derived from */c’/ by a rule [s] intervocally */c7 [o’] / */d/

{

Regular correspondences are found between the alveolar voiced stops of the Kuliak languages, irrespective of whether they are implosive or explosive. As this difference is phonemic in Ik, but apparently not so in Tepes and Nyang’i, the corresponding phoneme series is Ik /d/, /S/: Tp /d /: Ny /d/. Ik tud\at: Tp tud: Ny tud ‘five’ Ik dam: Tp dom: Ny dom ‘pot of earth* Ik 5 ok*: Tp dog: Ny dog ‘mucus’ Ik 5d5o; Ny dodo ‘sheep*

*jdzj A regularity of correspondences also seems to occur between Ik /dz/ and Tepes and Nyang’i 0 (zero). It is assumed that this series is derived from a proto-Kuliak phoneme */dz/ which has been lost in Tepes and Nyang’i. As only two examples of this correspondence have been found the reconstruction has to be considered very tentative. Ik dzib\er: Tp ib\e: Ny ibzr ‘axe’ Ik dzues\6s: Tp is\es ‘to steal’ The series Ik /dz/: T p 0 : Ny 0 seems to occur only morphemeinitially. There is, however, another series which is found only in morpheme-final position and therefore stands in complementary distribution which the former series: Ik /d /: Tp /s/1. lNo reflexes have yet been discovered in Nyang’i,

13

Ik bod\ok: Tp bzs ‘bark"(of tree)’ Ik k ’ad: Tp gas ‘to shoot’ As this series is phonetically similar to the former series23, it is also interpreted as a later development of */dz/. The hypothetical development of */dz/ can be summarized by the following rules:

*jdzj >

Ik Tp Ny

morphemeinitially /dz/

0 0

morphemefinally

m N

It has to be emphasized that for the series Ik /d/: Tp /s/, too, only two examples have been found. */g/ A series of correspondence is found between voiced velar stops, where Ik /g/ and / ’g/ correspond to Tepes /g/ and Nyang’i /g/. A proto-Kuliak phoneme */g/ is set up for this series. Ik gwah: Tp gwek: Ny gwezk ‘bird’ lk gwaits': Tp gwej: Ny gwej ‘giraffe’ Ik gaan: Ny gaan ‘bad’ Tp gud: Ny gud ‘navel’ This series does not occur preceding front vowels where the following series is found: Ik /j/: Tp jgj: Ny /j/. Ik jiik ‘always’: Tp giy\ek ‘daytime’: Ny jiy\ekw ‘day’ Ik jomzi Tp gim: Ny jim ‘earth, soil’ Tp gi: Ny j i ‘sun’ Tp gem: Ny jem 'to bite’ It is assumed that this series too. is derived from */g/. The replacement of the velar stops by a palatal stop in Ik and Nyang’i could then be interpreted as being connected with the palatal quality of the following vowels. There is a third series which can be related to */g/, i.e. Ik jgf: Tp0 : Ny0 . Only two examples have been found, however. Ik h ug-: Tp u-: Ny u- ‘to dig’ Ik ar\dgwan: Tp | dwan ‘moon’ 2It contains only alveolar phonemes and therefore shares the same place of articulation with the former series. 3Ik jom is derived from Proto-Kuliak *gIom (see 2.21).

14

This scries does not occur morpheme-initially whereas the other two series are found only at the beginning o f words or morphemes. It can therefore be assumed that */g/ has been lost in Tepes and Nyang’i non-initially, but was retained in Ik as /g/. w Ik has a phoneme /h/ which seems to occur only morphemeinitially and corresponds to 0 (zero) in Tepes and Nyang’i. For this series a proto-phoneme *jhj is reconstructed which is assumed to have been lost in Tepes and Nyang’i. Ik h&6: Tp a6: Ny ab ‘hot’ Ik hug-: Tp Ny u- ‘to dig’ Tepes, on the other hand, has a word-final glottal stop which corresponds to 0 in Ik and Nyang’i : Ik lo; Tp fo’: Ny fa ‘cattle’ Ik se: Tp s&’: Ny se ‘blood’ Ik /cwa: Tp oo\‘ Ny o ‘grass’ As the two series Ik jhj: Tp 0 : Ny 0 and lk 0 : Tp /’/: Ny 0 stand in complementary distribution and are phonetically sufficiently similar, they are tentatively grouped within one and the same proto-phoneme */h/. On the basis of this, it can he assumed that Proto-Kuliak had two sounds *[h] and * [ ?] occurring in comple­ mentary distribution; or, within a distinctive feature approach, it can be said that Proto-Kuliak had a rule of the kind

A series of correspondence is found between the voiced palatal stops of the Kuliak languages; Ik /j/ and / ’j/: Tp /j/: Ny /j/. For this series, a hypothetical phoneme */j/ is set up. Ik u’j a; Tp oj: Ny oj ‘sore’ Ik 7 Tp, Ny /g/ if followed by V 8/1 Ik /ny/ > Ik, Tp, Ny /o/ morpheme-initially > Ik /r/ Ny /ny/ elsewhere */ny/ A palatal nasal *jnyj is reconstructed for a series Ik /ny/: Tp /ny/: Ny /ny/. Ik yany: Tp ydny\u: Ny ydny\ u 'to open’ Ik n y e k Ny nyzg ‘hunger’ Tp nyem\u Ny: nyem\ur ‘duiker’ Tp muny\ay: Ny mimy\ay ‘yellow’ */p/ Ik /f/ corresponds regularly to /p/ in Tepes and Nyang’i. A Proto-Kuliak phoneme */p/ is set up, assuming that the voiceless bilabial stop has developed into a labio-dental fricative in Ik. Ik fad*: Tp pad: Ny pad ‘bitter’ Ik fen: Tp pen-\at: Ny pzn ‘fart’ Ik fo 'ja: Tp poj: Ny poj ‘to whistle’ 7 r/ A proto-phoneme */r/ is set up for a series Ik /r /: T p 0 : Ny /r/. This series neither occurs word-initially nor within monosyllabic morphemes. Ik dzib\er: Tp ib\i: Ny ibzr ‘axe’ Ik mor\dk: Tp mook: Ny mor\ok ‘throat’ Ik cfirtik: Tp ki\ uk: Ny kir\uk ‘bulls’ In word-initial position and in monosyllabic morphemes Tepes has a phoneme /r/ which corresponds to /r/ in Nyang’i and Ik. Tp rom\at: Ny room\dt ‘leaf’ Tp rib: Ny rib ‘sky’ Tp ri\dt: Ny riat (Driberg) ‘firewood’ 1This rule is not possible without a systematic comparison of the vowels; see 2.12, Variation o f vowels.

18

Ik 6ar: Tp 6ar ‘sour' Ik bzr: Tp her: Ny ber ‘to build, to mould* The two series, occurring in complementary distribution, can be said to be derived from the same Proto-Kuliak phoneme */r/. Tepes introduced a rule */r/ -> 0 in all except word-initial positions.

*N Another series Ik js}: Tp fsj: Ny /s/ seems to be derived from a proto-phoneme */s/. This series occurs both initially and finally but not word-internally between vowels. Ik bos: Tp bes: Ny bes ‘ear* Ik se: Tp js ’: Ny se ‘blood’ Ik sik’w*-: Tp stg: Ny sig\u ‘to sneeze’ Intervocally, the following series is found: Ik /r/: Tp fsj: Ny/s/. Ik morift a: Ny mos\id ‘bean’ Ik kzr'ok; Tp kos'ok ‘claw, finger’ Ik tor6b (Crazzolara): Ny tos\ob ‘chest’ This series, too, can be considered as being derived from ProtoKuliak VSA which through rhotacism developed in ter vocally to /r/ in Ik, */hl/ Ik and Nyang’i have a lateral continuant /hi/ which corresponds to /!/ in Tepes. A proto-phoneme */hl/ is set up for this series. */hl/ is assumed to have merged in Tepes with */l/. Ik yihl: Ny tjihl ‘hard, strong’ Ik hb>: Tp /o’: Ny hlo ‘cattle* Ik wahl: Tp wal ‘in front of’ Tp tek: Ny hlek ‘metals’ In word-final position */hl/ seems to have become /ihl/ in Nyang’i, unless preceded by */i/. Tp gel: Ny g&ihl ‘cow’ Tp rjal: Ny t)a\ihl ‘sharp’ w A phoneme */t/ is reconstructed for a series Ik ft/: Tp /t/: Ny /t/. Ik tud\at: Tp tud: Ny tud ‘five*

19

Ik ts'oot: Tp j\itjut: Ny juj\ut ‘to suck’ Ik tak\vG: Tp tak: Ny takw ‘to step on’ */w/ All three Kuliak languages have labio-valar semi-vowels which correspond regularly to each other. Underlying the series Ik /w /: Tp /w/: Ny /w/ is a proto-phoneme */w/ Ik wet; Tp wet: Ny we:t ‘to drink’ Ik wahl; Tp wal: ‘in front of’ Tp nowd; Ny nowje ‘four’ Tp tow\d ‘to forbid’: Ny tow\ec ‘to refuse’ There is another series Ik /w /: Tp 0 : Ny /w/ which only occurs word-finally following velar stops. As the former series is not found in this environment, both series can be said to be derived from the same hypothetical proto-phoneme */w/. In Tepes a rule of the following form was introduced: $ '— vocalic '— contin." — conson. _+ grave

0 1

— diffuse _+ grave

Ik ekwa: Tp ek; Ny ekw ‘eye’ Ik takwB: Tp tak: Ny takw 'to step on’ Ik agwjeh a; Tp ag: Ny agw ‘in inside’ There are a few instances where *(\vj seems to have been lost in Ik, in the same environment as in Tepes, e.g, Ik tak'a ‘shoe’ (Crazzolara): Tp teg: Ny tegw ‘foot, leg* Ik jiik ‘always’; Tp giy\$k ‘daytime’: Ny jiyjekw ‘day’

*/y/ A proto-Kuliak phoneme */y/ is reconstructed for a series Ik 0 : Tp /y/: N y0 . Ik ma: Tp may: Ny ma ‘sick’ Ik edA: Tp yed: Ny edek ‘name’ Tp yey: Ny ye ‘meats’ This series occurs only word-initially and word-finally. Wordinternally, .another series Ik fyf : Tp /y/: Ny fyf is found which is likely to be derived from the same proto-phoneme. Ik iy\es: Tp yes 'to know’

20

Tp diy\Sn: Ny diy\an ‘to speak’ Tp in - iy\on : Ny iy\on ‘three’ Ik ma pi. mayak\at: Tp may ‘sick’ The development of Proto-Kuliak */y/ would then imply that Ik and Nyang’i introduced a rule */y/ 0 if preceded or followed by a word-boundary This rule must have been introduced independently in each of the two languages (see 3.2).

We may assume that Proto-Kuliak also had a voiced alveolar fricative */z/ which is reflected as jzj in Ek and /s/ in Nyang’i. No reflexes are found in Tepes. Ik zik: Ny sik- ‘to tie, to fasten’ Ik zlz ‘fat*: Ny s\fs- ‘to become fat* There are a few other series of consonantal correspondences which, however, are not interpreted as being derived from simple proto-consonants but rather from combinations of Proto-Kuliak phonemes.

*/ky/ Ik has an alveo-palatal phoneme /c/ which corresponds regularly to /k/ in Tepes and Nyang’i. Ik cn|e; Tp kwe\- Ny kwe ‘water* Ik curtih: Tp ki\iik: Ny kir\uk ‘bulls* Ik c&m\on: Tp k m\dn: Ny ksm\6n ‘war’ It will be assumed that this series goes back to a combination of the proto-phonemes */k/ and */y/, and that the palatal articulation ""of */y/ was lost in Tepes and Nyang’i, but was responsible for the development */ky/ > Ik /c/. The development */ky/ > Ik /c/ must have been preceded by the development */c/ > Ik /ts/, otherwise the two Proto-Kuliak segments would have merged in Ik.1 In word-final position, */ky/ possibly developed into Jhj in lk,

1Whether the vacuum left in the phonological system of Ik by the shift */c/ > As/ was responsible for the subsequent shift * /k y /> /c/ remains to be investigated/.

21

C-EMlTiSKA SEMINARlET

as the following cognates suggest: Ik gwah pi. gwaicik: Tp gwek: Ny gm&k ‘bird’ */rr/ As has been indicated above, Proto-Kuliak */r/ was lost in Tepes except in word-initial position and within monosyllabic morphemes. There is however a series Ik /r/; Tp /r/: Ny /r/ which occurs in this environment, e,g. Ik ar\i&: Tp ar\i ‘guts’ Tp oru ‘to pour on the ground’ : Ny or\u ‘to pour’ Tp ibr\an: Ny iber\£n ‘axes’ It is tentatively assumed that this series is derived from a redu­ plicated */r/. */kw/ It seems that the phoneme combination */kw/ has become /w/ in Tepes in word-initial position. There is, however, only one example available: Ik kwats: Tp wac: Ny kwac\ik ‘urine’

22

2,12

Vowels

A Proto-Kuliak phoneme */a/ is set up for a series Ik /a/. Tp /a/. Ny /a/. Ik ah*: Tp dk: Ny ak ‘mouth* Ik ats: Tp dc: Ny dc ‘to come’ Ik kan: Tp k&n: Ny kan ‘to lick’ Another series Ik /a/: Tp /e/: Ny /e/ also seems to be derived from Proto-Kuliak */a/Ik fl»a: Tp eu: Ny e\u ‘village, family, home’ Ik mau: Ny mjeu ‘lion’ Ik tsak\dd: Tp cegddt *cekudat): Ny cekud fare-stick The shift */a/> T p, Ny /e/ probably occurred whenever the follow­ ing syllable contained a high vowel, i.e. a vowel having the feature [4-diffuse]. This development can be described by a rule like . *4- vocalic "4- compact ‘— compact 1 i _ — conson, 4- grave — grave 4- diffuse — flat where C = any consonant

J

A similar rule can account for another development ProtoKuliak */a/ > Ik /a/: Tp /s/: Ny /e/. Ik da: Ny dz ‘good, nice’ ' Ik mayak\dt: Tp mzy ‘sick (pi.)’ Ik takda ‘shoe’ (Cra): Tp tzg: Ny tzgw ‘foot, leg’ This series occurs if */a/ is followed by a semi-vowel in the same syllable and does not bear stress2. In stressed position, the rule

23

*/a/ > Tp, Ny /e/ does not apply, e.g. Ik ma: Tp m&y: Ny ma ‘sick’ ( < *mjay) Tp \eak: Ny ejakw ‘to herd animals’ ( < *e }jakw) Furthermore, there is a series Ik /a/: Tp /a/: Ny /e/ which can also be related to the proto-phoneme */a/. Ik e\dg: Ny iarj-ac ‘to rest* Tp im\an: Ny im\zn ‘black* Tp nowa’: Ny na>v|e ‘four’ Tp jow\a\- Ny jow |e ‘guts (pi.)* This series occurs if the proto-vowel preceding it is either mid or high, i.e. has the feature [—compact].1 */«/ Another Proto-Kuliak phoneme, */e/, is reconstructed for a series Ik /e/: Tp /e/: Ny jzj. Ik zb0-: Tp eh: Ny zb ‘horn’ Ik bzr: Tp her: Ny hzr ‘to build’ Ik wzt~: Tp wet-; Ny wszt- ‘to drink’

*/e/ A separate proto-phoneme */e/ is set up for a series Ik /e/: Tp /e/: Ny /e/. Ik id 0-: Tp yed: Ny edek ‘name* Ik e\at: Tp \eat ‘to increase’ Ik e\ak'a: Tp e\ag: Ny e\ag ‘far’ */?/ The series Ik ji ( : Tp /e /: Ny /e/ is considered to be derived from a Proto-Kuliak phoneme other than *(z( or */«/• Another hypo­ thetical proto-phoneme */$/ is set up, which is distinguished from */e/ by a feature [ + tense], whereas */e/ has f—tense]. Ik /cf/a: Tp ej ‘ground’: Ny ej ‘land’ Ik \ibit: Ny ebita ‘to plant*

1There are however cases where the rule */a/ > Ny /s / /[—compact] (C)— does not apply, e.g. Tp munyidJ: Ny munyjdo ‘yellow’ We still need to investigate to what extent this rule applies.

24

Ik -ik: Tp -ek: Ny -ek nominal plural suffix Ik bits\in Tp bec\in: Ny bec\tn ‘leftside’ ■w From this series, another series Ik /e/: Tp (if: Ny /i/ is to be dis­ tinguished. It is assumed to be derived from Proto-Kuliak */I/ which differs from */e/ by the features f + diffuse 1 whereas */e/ L - tense

J

Ik Ik Ik Ik

ets*: Tp ij: Ny ij\6o ‘excrement(s)’ 'je'j-: Tp jij ‘to remain’ ar\&&: Tp or (/‘guts’ em\(n: Tp \imin 'to pull’

*N A phoneme */i/ is reconstructed for a series Ik (if: Tp /i/: Ny /i/. *jij differs from */I/ by its feature [ + tense]. Ik id*: Tp id: Ny id ‘milk’ Ik ik*: Tp ik: Ny ik ‘head’ Ik dzib\er: Tp ib\i: Ny iber ‘axe’ W A series of regular correspondences Ik (of: Tp (of: Ny (of is taken to be derived trom another Proto-Kuliak phoneme *fof '+ compact which has the distinctive features + grave _+ flat Ik Sok?: Tp dog: Ny dog ‘mucus’ Ik hlo: Tp lo': Ny hlo ‘cattle’ Ik ok*: Tp ok: Ny ok\dt ‘bone’ */o/ A proto-phoneme */o/ is reconstructed for a series Ik /o/; Tp (of: Ny (oj which differs from */o/ bv the feature [—compact]. Ik dom: Tp dom: Ny dom ‘pot of earth’ Ik idrdb ‘chest’ (Cra): Ny tos\ob ‘front part of body’

25

Tp or: Ny or ‘hole’ Tp od: Ny od ‘road’ W From */°/ another proto-phoneme */