PHONOLOGICAL AND MORPHOLOGICAL NATIVIZATION OF …

PHONOLOGICAL AND MORPHOLOGICAL NATIVIZATION OF ENGLISH NOUNS

BORROWED INTO EKEGUSII A CONSTRAINT-BASED APPROACH

ANYONA GEORGE MORARA

BA (UNIVERSITY OF NAIROBI) PGDE (KENYATTA UNIVERSITY) MA

(EGERTON UNIVERSITY)

A THESIS SUBMITTED TO THE SCHOOL OF POST-GRADUATE STUDIES IN

PARTIAL FULFILLMENT OF THE REQUIREMENTS OF THE DEGREE OF

DOCTOR OF PHILOSOPHY OF THE FACULTY OF ARTS AND SOCIAL SCIENCES

DEPARTMENT OF LINGUISTICS LANGUAGES AND LITERATURE

KISII UNIVERSITY

NOVEMBER 2017

DECLARATION

DECLARATION BY THE CANDIDATE

This thesis is my original work and it has not been submitted in this or any other university

known to me

Signhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip Datehelliphelliphelliphelliphelliphelliphelliphelliphellip

George Morara Anyona

DAS13600102014

DECLARATION BY THE SUPERVISORS

This thesis has been submitted with our approval as university supervisors

Sign helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip Datehelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

Dr David O Ongarora PhD

Department of Linguistics Maseno University

Dr Evans G Mecha PhD

Department of Lit Lang amp Ling Kisii University

PLAGIARISM DECLARATION

DECLARATION BY STUDENT

i I declare I have read and understood Kisii University Postgraduate Examination Rules

and Regulations and other documents concerning academic dishonesty

ii I do understand that ignorance of these rules and regulations is not an excuse for a

violation of the said rules

iii If I have any questions or doubts I realize that it is my respondibility to keep seeking an

answer until I understand

iv I understand I must do my own work

v I also understand that if I commit any act of academic dishonesdty like plagiarism my

thesisproject can be assigned a fail grade (ldquoFrdquo)

vi I further understand I may be suspended or expelled from the university for academic

dishonesty

Name George Morara Anyona Signaturehelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

Reg No DAS13600102014 Datehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

DECLARATION BY SUPERVISOR(S)

i Iwe declare that this thesis has been submitted to plagiarism detection service

ii The thesis contains less than 20 of plagiarized work

iii Iwe hereby give consent for making

1 Name Dr David O Ongarora PhD Signaturehelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

Affiliationhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip Datehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

2 Name Dr Evans G Mecha PhD Signaturehelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

DECLARATION OF NUMBER OF WORDS

Name of Candidate GEORGE MORARA ANYONA Adm No DAS13600102014

Faculty ARTS AND SOCIAL SCIENCES

Department LINGUISTICS LANGUAGES AND LITERATURE

Thesis Title PHONOLOGICAL AND MORPHOLOGICAL NATIVIZATION OF ENGLISH

NOUNS BORROWED INTO EKEGUSII A CONSTRAINT-BASED

APPROACH

I confirm that the word length of

1) The thesis including footnotes is 64 004 2) the bibliography is 3 972

And if applicable 3) the appendices are 3 219

I also declare the electronic version is identical to the final hard bound copy of the thesis and

corresponds with those on which the examiners based their recommendation fro the award of the

degree

Signedhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip Datehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

(Candidate)

I confirm that the thesis submitted by the above-named candidate complies with the relevant

word length specified in the School of Postgraduate and Commission of University Education

regulations for the Masters and PhD Degrees

SignedhelliphelliphelliphelliphelliphelliphelliphelliphellipEmailhelliphelliphelliphelliphelliphelliphellipTelhelliphelliphelliphelliphelliphellipDatehelliphelliphelliphelliphellip

(Supervisor 1)

(Supervisor 2)

COPYRIGHT

All rights are reserved No aprt of this thesis or information herein may be reproduced stored in

a retrieval system or transmitted in any form or by any means electronic mechanical

photocopying recording or otherwise without the prior written permission of the author of Kisii

University on that behalf

copy 2017 Anyona Morara George

ABSTRACT

This study examines the phonology and morphology of English nouns borrowed into EkeGusiiwithin Optimality Theory This theory provides that the well formedness in natural languages isconstraint governed Thus even borrowed lexical items obey these constraints Constraints areuniversal and are ranked on a language specific order The focus of this study is to investigate thephonological and morphological adjustments that English nouns borrowed into EkeGusiiundergo as they harmonize themselves with EkeGusii phonological and morphological systemsas they adapt EkeGusii constraints ranking order The study analyzes segmental phonotactic andsuprasegmental adaptations that English nouns borrowed into EkeGusii undergo in the process ofbeing nativized Further the study analyzes the affixation processes that characterize thesenouns The study analyses nominal prefixation augmentation and classification The study wasguided by the following objectives to describe the phonological and morphological structures ofEkeGusii and English nouns to analyze the phonological adjustments that English nounsborrowed into EkeGusii undergo during nativization and to analyze the morphological changesthat the English noun borrowed into EkeGusii undergo during nativization The study adoptedthe descriptive and the explanatory research designs Data was collected from a sample of 13interviewees using semi-structured interviews The interviewees were purposively sampledbased on the semantic domains of the nouns collected All the 349 English nouns borrowed intoEkeGusii colleccted from the field constituted the sample size of the study secondary data wasused to describe the phonological and morphological strucures of EkeGusii in response toquestion 1 Library study and researcherrsquos intuition were the sources of the secondary data Totest the validity and reliability of the research instrument pre-testing was carried out andappropriate adjustments and corrections made on the instument The study yielded the followingresults The first objective describes the phonological and morphological structures of Englishand EkeGusii It was established that Phonologically the acoustic nature of the vowels in thetwo languages differ significantly EkeGusii syllable structure is (V)CV while that of English is(C)V(C) and EkeGusii is characterized by a number of phonological processes not found inEnglish Morphologically EkeGusii nouns unlike those of English are grouped into classesdetermined by the semantics of the root of the noun and affixation in the two languages isdifferent The second objective analyzed the phonological adjustments that English nounsborrowed into EkeGusii undergo It was established that phonemes of English not present inEkeGusii are substituted for those present in EkeGusii the foreign syllable structure fromEnglish is re-syllabified to EkeGusii syllable structure and English stress is tonemized inEkeGusiiThe third objective analyzed the morphological changes that English nouns borrowedinto EkeGusii undergo during nativization It was found that English nouns enter EkeGusiinominal classes which are semantically determined nativization of the English nouns borrowedinto EkeGusii is characterized by the pre-prefix and semantics plays a significant role inmorphological nativization of English nouns in EkeGusii This study has establisshed thatOptimality Theory can successively analyse English nouns borrowed into EkeGusii and that thenouns obey EkeGusii constraint ranking It is hoped that this study would contribute to thetheoretical understanding of borrowed word phonology and morphology of EkeGusii inparticular and that of Bantu languages in general

DEDICATION

To Moraa my spouse and my children Anyona Nyantari Nyaboke Kerubo and Mokeira

ACKNOWLEDGEMENT

Many hands helped in coming up with this dissertation but given the limitations of space I may

not mention all of them nor will I record all their individual contributions However the

following people and institutions deserve special mentioning and recognition

Much credit goes to my supervisors Dr Evans Gesura Mecha and Dr David Ogoti Ongarora

for their valuable and unending advice guidance and direction right from the time of topic

selection and refining and indeed to the final preparation of this dissertation You were a

blessing from God Thank you

My sincere appreciations go to Professor Augustine Agwele of Texas State University-USA (an

adjunct professor in KSU) firstly for his scholarly advice and mentorship secondly for

teaching me scholary patience and thirdly and most importantly for taking his time to read

critique and edit this work Thanks a lot prof

I also take this opportunity to thank Dr Nilson Opande former COD Department of Ling

Lang and Lit now director Board of Undergraduate Studies Dr Barasa Margaret former

COD Ling Lang and Lit now current Dean FASS Kisii University for their encouragement

and efforts of all manner of kinds to ensure that this work is completed within time You are not

only my senior colleagues and mentors but also my dear friends as well

As it is to every writer of a work of this nature I owe a debt to all my teachers from pre-school

to this level all my colleagues and my dear students to you all I gratefully acknowledge my

deep indebtness

Much credit also go to the thirteen (13) great men and women of Nyagaachi village Echoro sub-

location of Nyamira County led by their able assistant chief Mr David Orina and village elder

Mr James Aricha for readily and willingly availing themselves for interviewing and thus

providing the data that was analyzed in this study Your eagerness and willingness to participate

in this study was simply amazing To all of you I say God bless you

I would also like to extend my appreciations to all my course mates Magutu Omari Nyoteyo

and Omosa for their unending advice and input during course work and indeed during the

preparation of this thesis To you all I say yes we can

I will not be fair if I do not mention the ladies who spent several hours typing and formatting this

work during its preparation To Regina Obonyo and Rhoda Makori I say thank you and God

bless you for your efforts patience and time

I will not end these aknowlegements without mentioning the following institutions firstly Kisii

University for giving me the opportunity to be one of its pioneering PhD students in Linguistics

secondly the department of Ling Lang and Lit for providing human and material support

required to undertake the programme most notably the language laboratory thirdly the

Counnty government of Nyamira for not only allowing me to conduct research in Couny but

also providing to me all the information requested for and finally the Government of Keny

through NACOSTI for permitting me to carry out this work I am sincerely grateful to all these

institutions

Finally and by no means not least to my dear wife Mary my daughters Laura Esther and

Joyline my sons Edwin and Robinson my mother Nyaboke and all my sisters and brothers I

say a big thank you for your time constant and continuous encouragement and support

understanding and patience during the painful period of preparing this thesis You gave me a

peace of mind required in doing this kind of work God bless you all

TABLE OF CONTENTS

DECLARATIONii

PLAGIARISM DECLARATIONiii

DECLARATION OF NUMBER OF WORDSiv

COPYRIGHTv

ABSTRACTvi

DEDICATIONvii

ACKNOWLEDGEMENTviii

TABLE OF CONTENTSx

LIST OF ABBREVIATIONS SYMBOLS TABLES TABLEAUX FIGURES AND CHARTS

CHAPTER ONE

INTRODUCTION1

10 Background to the study1

11 Statement of the problem5

12 Objectives of the research6

13 Research questions6

14 Justification and significance of the study6

15 Scope and limitation9

CHAPTER TWO

LITERATURE REVIEW AND THEORETICAL FRAMEWORK12

20 Introduction12

21 EkeGusii phonological and morphological structures14

211 EkeGusii phonological structure14

212 EkeGusii morphological structure18

22 Phonological nativization19

23 Morphological nativization25

24 Theoretical framework28

241 Tenets of Optimality Theory28

242 Optimality Theory versus other Generative Approaches35

CHAPTER THREE

RESEARCH METHODOLOGY40

30 Introduction40

31 Research design40

32 Research site41

33 Study populations41

34 Sample size and sampling techniques42

35 Data collection procedures43

36 Data analysis43

361 Procedure data analysis44

37 Ethical considerations45

CHAPTER FOUR

DATA PRESENTATION ANALYSIS RESULTS AND DISCUSSIONS46

40 Introduction46

41 Phonological and morphological structures of EkeGusii and English46

411 EkeGusii and English vowel systems47

4111EkeGusii vowel harmony and disharmony59

412 EkeGusii consonant system in comparison with English63

4121 EkeGusii consonantal processes70

41211 Voice dissimilation (Dahlrsquos Law)70

41212 Prenasalisation and nasal homorganisation72

41213 Declusterization of nasal consonant (NC) and consonant glide (CG) clusters73

41214 Defricativisation82

41215 EkeGusii Nasal re-syllabification83

413 Comparison of EkeGusii and English syllable structures89

4131 EkeGusii and English phonotactics102

41311 English word initial phoneme sequences103

41312 English word final phoneme sequences105

414 Comparison of EkeGusii toneme structure and English stress107

4141 EkeGusii tone structure107

41411 Underlying and surface tones in EkeGusii109

41412 Contour tones in EkeGusii110

41413 Tone floating in EkeGusii112

415 Comparison of EkeGusii and English morphological structures114

4151 Comparison of EkeGusii and English noun systems115

4152 Morphosyntactic classes and prefixes in EkeGusii noun116

41521 The pre-prefix or augment121

41522 The Prefix125

41523 EkeGusii noun class roots127

41524 Class 1 omo- and 2 aba-128

41525 Classes 3 omo- and 4 eme-128

41526 Classes 9 e- 10 chi- 9 (a) en- 10 (a) chin-130

41527 Classes 5 eri-rii- 6 ama- 7 eke- 8 ebi- 12 aka- and 14 obo-131

42 Phonological nativization of English nouns in EkeGusii135

421 Segmental nativization136

4211 Nativization of vowel segments137

42111 Nativization of English pure (monophthong) vowels138

42112 Nativization of English diphthongs157

42113 Nativization of English triphthongs166

4212 Nativization of English consonants173

42121 Nativization of English f and v174

42122 Nativization of English eth and 179

42123 Nativization of English l183

42124 Nativization of English z187

42125 Nativization of English int and ʒ189

42126 Nativization of English g in EkeGusii192

422 Phonotactic nativization195

4221 Nativization by vowel epenthesis200

42211 Nativization of complex syllable margins201

42212 Nativization of closed syllables205

423 Supra-segmental nativization210

424 Nativization by phonological processes216

4241 Nativization by voice dissimilation (Dahlrsquos Law)216

42421 nativization by defricativization225

42422 Spirantizationfricativization of the stops230

42423 Nativization by bilabialization of labiodentals233

4243 Nativization of nasal consonant clusters (NC) by pre-nasalization and homorganization

4244 Nativization by vowel height harmonization and disharmonization239

431 Nativization by nominal classification243

4311 The role of semantics in morphological nativization of English nouns in English252

4312 Nativization into classes 1 and 2 Nouns262

4313 Nativization of English loans into EkeGusii in classes 1b mo- and 2b [ba-264

4314 Nativization into classes 3 and 4265

4315 Nativization in classes 5 eri-rii- 6 ama- 7 eke- 8 ebi- 12 aka- and 14

obo-265

432 Prefixation and pre-prefixation268

4321 Nativization by prefixation269

4322 Nativization by pre-prefixation272

CHAPTER FIVE

SUMMARY CONCLUSIONS AND RECOMMENDATIONS276

51 Summary276

52 Conclusions281

53 Recommendations282

REFERENCES285

APPENDICES307

Appendix I Interviewee profile form307

Appendix II Interviewee consent form308

Appendix III Research instrument semi-structured interview guide309

Appendix IV Raw data recordingtranscription form312

Appendix V Raw data313

Appendix VI Illustrations of phonological nativization327

Appendix VII Illustrations of morphological nativization328

Appendix VIII Enchoro sub-location villages329

Appendix IX Research Clearance Permit330

Appendix X Research Authorization Letter331

Appendix XI Memo from the Registrar Research and Extension Kisii University to NACOSTI

Appendix XII Plagiarism Report333

Appendix XIII Publication336

LIST OF ABBREVIATIONS SYMBOLS TABLES TABLEAUX FIGURES AND

CHARTS

ABBREVIATIONS

AFX affix

APP applicative

AUG a pre-prefix (a vowel)

C consonant

CON constraints (in the Optimality theory)

EVAL evaluator

FV final vowel

GEN generator

OT Optimality Theory

IND indicative mood

L left edge of a word

N nucleus (of a syllable)

NC nasal Consonant

O onset consonant (of a syllable)

OM object Marker

P nominal Person

PL plural

SG singular

SM subject Marker

1 3 SG class one noun in the third person singular

7 2 PL class seven noun in the second person plural

R right edge of a word

RT root of a word

TNS tense

V vowel

H high for tonal tier high tone

HTS high tone spans

L O low for tonal tier

ATR advanced tongue root

RTR retracted tongue root

UF underlying form of a language

SF surface form of a language

OCP obligatory contour principle

TBU tone bearing unit

FAITH C an Optimality Theory constraint that demands that consonants in the input

be the same as the consonants in the outputFAITH V an Optimality Theory constraint that demands that vowels in the input be

the same as the vowels in the outputIDENT-IO (F) the specification that the features of an input segment must be preserved

in its output correspondentINDENT-IO (place) the specification that place of articulation of an input segment must be

preserved in its output correspondent

VOI voiced obstruents are not allowed

VTV voiceless obstruents are banned in intervocalic positions

MAX- IO input segments must have output correspondents (no deletion)MAX-V input vowels must have output correspondents(no deletion)CODA syllables are open ( must not have codas)ONSET syllables must have onsets

DEP no epenthesis or addition of either a vowel or consonant

OCP(dorstop) a markedness constraint which bans sequence of voices dorsal stop

OCP(V) a markedness constraint which bans sequence of voiced consonants

COPLEX(C) a constraint which bans complex consonants

COPLEX(V) a constraint which bans a string of vowels

BACK V back vowels are not allowed

CENTRAL V central vowel are not allowed

LAX V lax vowels are not allowed

HIGH V high vowels are not allowed

AGREE V vowels in a word should agree

AGREE C consonants in a word should agree

STRIDENT (FRIC) strident fricates are not allowed

INTERD (FRIC) interdental fricatives are not allowed

LATERAL (C) lateral consonants are not allowed

TRILL (C) trill consonants are not allowed

VOICEOBS voiced obstruents are not allowed

NONANTERCORFRIC non-anterior coronal fricatives are not allowed

ASSIM (F) assimilation of features are not allowed in a given domain

UNIFORMITY prohibits feature fusing

FLOAT markedness constraint which prohibits floating of tones

SPECIFY a constraint which demands that a tone bearing unit must have a

correspondent tone

MAX IO (MORPH) morphological faithfulness constraint which demands that an input

morpheme must have an output correspondent no change

DEP IO (MORPH) a morphological faithfulness constraint which demands that there should

be no epenthesis of a morpheme

(ALIGN(AFX R RT L)) a morphological alignment constraint which demands that the right

edge of an affix should be aligned to the left edge of a root (a prefix)

(ALIGN(AFX L RTR)) a morphological alignment constraint which demands that the left edge

of an affix should be aligned to the right edge of a root (a suffix)

STRPRES a faithfulness constraint which demands that a structure in the input should

be preserved in the output no structure change

SYMBOLS

Slashes enclose phonemic forms

[ ] Square brackets enclose phonetic forms

Braces enclose morphemes

| | Pipes enclose underlying forms

rarr Arrow a phonological form realized as

gt Morphological form meaning realized as

gtgt Domination (constraint) in ranking

σ Syllable node

micro Syllabic molar

Constraint violation

Optimal candidate

Fatal violation

Unbroken association line indicating prelinking in Autosegmental

Phonology

Broken association lines indicating linking in Autosegmental phonology

ndash Tone marker for level tone` Tone marker (low tone)acute Tone marker (high tone) syllable markerOslash A nonexistent segment to be substituted or inserted for another

TABLES

Table 1 Some of the OT constraints from the universal sethelliphelliphelliphelliphelliphelliphelliphelliphellip 31

Table 2 F1 and F2 formant frequencies in hertz (Hz) of EkeGusii vowelshelliphelliphellip 51

Table 3 F1 and F2 formant frequencies in hertz (Hz) of EkeGusii vowels[i e ͻ u]hellip 56

Table 4 F1 and F2 formant frequencies in hertz (Hz) of the English vowels [i e ͻ u]helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 56

Table 5 Typology of syllable shapeshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 89

Table 6 EkeGusii prefixes and their stems semantic determinantshelliphelliphelliphelliphellip 118

TABLEAUX

Tableau 21 Modern English realization of the input intip helliphelliphelliphelliphelliphelliphelliphellip34

Tableau 22 Comparative tableau for Yawelmani input Ɂilk-hinhelliphelliphelliphelliphelliphellip37

Tableau 23 Violation tableau for Yawelmani input Ɂilk-hinhelliphelliphelliphelliphelliphelliphellip 38

Tableau 41 English output of the input krɪsmǝshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 141Tableau 42 EkeGusii output of the input kirismasihelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 142Tableau 43 English input and output of intɜt helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 148Tableau 44 EkeGusii output of the input esati helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 148Tableau 45 EkeGusii output of the input ekerasihelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 150Tableau 46 English input and output of intɜt helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 150Tableau 47 English input and output of sɪneɪthelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 153Tableau 48 EkeGusii output of the input esenetihelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 154Tableau 49 EKeGusii realization of the input erokesenihelliphelliphelliphelliphelliphelliphelliphelliphellip160Tableau 410 English output of the input lǝʊkeɪintnhelliphelliphelliphelliphelliphelliphellip 160Tableau 411 EkeGusii realization of the input wajahelliphelliphelliphelliphelliphellip 171Tableau 412 English input and output of waɪǝhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 171Tableau 413 EkeGusii output of the input euritintihelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 177Tableau 414 English input and output of frɪdʒhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 177Tableau 415 Tonga realization of the English input frɪdʒhelliphelliphelliphelliphelliphelliphelliphellip 179Tableau 416 English output of the input Өɜməshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 182Tableau 417 EkeGusii output of the the input etamosihelliphelliphelliphelliphelliphellip helliphelliphellip182Tableau 418 EkeGusii output of the the input ekerasihelliphelliphelliphelliphelliphelliphelliphelliphelliphellip186Tableau 419 KiKamba realization of the English input skulhelliphelliphelliphelliphelliphellip 186Tableau 420 EkeGusii output of the the input esirohelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 188Tableau 421 English output of the input zirəʊhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 189Tableau 422 EkeGusii output of the input esati helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 191Tableau 423 English output of the input intɜt helliphelliphelliphelliphellip helliphelliphelliphelliphelliphelliphellip 191Tableau 424 EkeGusii output for the input epaγihelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 194Tableau 425 English output for the input baeligghelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 194

Tableau 426 EkeGusii output and input of mamahelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 198

Tableau 427 English output for the input ǝʊhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 199

Tableau 428 English output of the input teɪkhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 199

Tableau 429 English output of the input steɪnshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 199

Tableau 430 English output of the input skulhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 203

Tableau 431 EkeGusii output of the input sukuruhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 203

Tableau 432 English output of the input tͻtinthelliphelliphelliphelliphelliphelliphelliphelliphelliphellip helliphelliphellip207

Tableau 433 EkeGusii output for the input [tͻͻtinti]helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 207

Tableau 434 Lanakel output of the input [t-n-ak-ol]helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 208

Tableau 435 Lanakel output of the input [no-n-koma-i]helliphelliphelliphelliphelliphelliphelliphelliphellip 210

Tableau 436 EkeGusii output of the input oraŋgeti helliphelliphelliphelliphellip 215

Tableau 437 EkeGusii output of the input okokorohelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 219

Tableau 438 English output of the input input baeligŋkhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip222

Tableau 439 EkeGusii output of the English input eeŋgihelliphelliphelliphelliphelliphelliphelliphellip222

Tableau 440 EkeGusii output of the input kanihelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip229

Tableau 441 EkeGusii output for the input etaγisihelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip232

Tableau 442 EkeGusii output of the input arandahelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 234

Tableau 443 EkeGusii output for the input aranda (unit)helliphelliphelliphelliphelliphelliphelliphellip 237

Tableau 444 EkeGusii output of the input εntεrεahelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 240

Tableau 445 EkeGusii output for the neutral input nkookohelliphelliphelliphelliphelliphelliphellip 244

Tableau 446 EkeGusii output for the input nkooko helliphelliphelliphelliphelliphelliphelliphellip 245

Tableau 447 English output for the singular form input boyhelliphelliphelliphelliphelliphellip 246

Tableau 448 English output of the input boy-shelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 246

Tableau 449 English output the plural input blanket-shelliphelliphelliphelliphelliphelliphelliphellip248

Tableau 450 EkeGusii output of the English plural input blanket-shelliphelliphellip 248

Tableau 451 English output of the input tractor-shelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 254

Tableau 452 EkeGusii output of the English input tractor-shelliphelliphelliphelliphelliphellip 255

Tableau 453 EkeGusii output for input omo-tehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 257

Tableau 454 English output for the input records-shelliphelliphelliphelliphelliphelliphelliphelliphellip 258

Tableau 455 English output of the input recordhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 259

Tableau 456 EkeGusii output of the English singular input record-shelliphellip 260

Tableau 457 Tableau (457) EkeGusii output of the English input record-shelliphellip 261

Tableau 458 EkeGusii output of the English input omo-chiibuhelliphelliphelliphelliphellip 263

Tableau 459 English output of the input chiefhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip264

Tableau 460 EkeGusii output of the English input blankethelliphelliphelliphelliphelliphelliphellip 270

Tableau 461 EkeGusii output of the English input gaβanahelliphelliphelliphelliphelliphelliphellip 275

CHARTS Page

Chart 1 EkeGusii vowelshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 47Chart 2 EkeGusii vowel diagramhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 48Chart 3 EkeGusii consonantshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 64Chart 4 EkeGusii consonant inventoryhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 66Chart 5 The English consonantshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 68Chart 6 Description of English ʌ and EkeGusii ahelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 146Chart 7 Production of the English diphthong [aɪ]helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 163Chart 8 Production of EkeGusii nativized form [ae] from the English

Diphthong [aɪ]helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 163

FIGURES Page

Figure 1 Process of candidate elimination in OThelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip30

Figure 2 Process of OT realization of outputhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip44Figure 3 Spectrograms of EkeGusii vowels [a e ε i ͻ o u] collected from a

Native speakerhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 50Figure 4 Plot of F1 against (F2) formant frequencies in hertz (Hz) of EkeGusii

vowelshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip51

Figure 5 Spectrograms of 8 British English vowelshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 57

Figure 6 English diphthongshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip58

Fifure 7 Articulation of complex glides in EkeGusiihelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 76

Figure 8 EkeGusii consonant glide syllabic nodeshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 77

Figure 9 Nasal consonant syllabification in EkeGusiihelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 79

Figure 10 EkeGusii vowel lengtheninghelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 81

Figure 11 EkeGusii compensatory lengtheninghelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 81

Figure 12 EkeGusii de-linking rulehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip85

Figure 13 EkeGusii nasal re-syllabification rulehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 86

Figure 14 Nasal delinking leading to lasal syllabificationhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 87

Figure 15 Nasal consonant cluster forming part of a syllablehelliphelliphelliphelliphelliphelliphelliphelliphellip 87

Figure 16 EkeGusii syllable nodes for ketiihelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 100

Figure 17 EkeGusii syllable nodes for [omote] and [aaa]helliphelliphelliphelliphelliphelliphelliphelliphelliphellip 101

Figure 18 EkeGusii lexical contrasts due to tone structurehelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 108

Figure 20 EkeGusii tone markinghelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 110

Figure 21 Mende contour toneshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 111

Figure 22 EkeGusii LH contour tonehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 111

Figure 23 Ekegusii level tone 112

Figure 24 EkeGusii floating tone helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 112

Figure 25 EkeGusii prefix and pre-prefix syllable nodes helliphelliphelliphelliphelliphelliphelliphelliphelliphellip 121

Figure 26 Substitution of the English ɪ for EkeGusii [i]helliphelliphelliphelliphelliphelliphelliphelliphelliphellip 139

Figure 27 EkeGusii nativization of the English central vowels helliphelliphelliphelliphelliphelliphelliphellip 155Figure 28 Substitution of the English ǝɪǝ for EkeGusii [aja]helliphelliphelliphelliphelliphelliphelliphelliphellip 167Figure 29 Substitution of the English f for EkeGusii [β]helliphelliphelliphelliphelliphelliphelliphelliphelliphellip 176Figure 30 Substitution of the English Ө for EkeGusii [t]helliphelliphelliphelliphelliphelliphelliphelliphelliphellip 180Figure 31 Substitution of the English l for EkeGusii [r]helliphelliphelliphelliphelliphelliphelliphelliphelliphellip 183Figure 32 Substitution of the English z for EkeGusii [s]helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip187Figure 33 Substitution of the English int for EkeGusii [s]helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 190Figure 34 Substitution of the English g for EkeGusii [γ]helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip193Figure 35 EkeGusii syllable types helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 197

Figure 36 English syllable typeshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 197

Figure 37 Nativization by anaptyxishelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 202

Figure 38 Increase of number of Syllables in nativized word forms due to anaptyxishelliphelliphellip204

Figure 39 Opening of the English closed syllables by paragogic epenthesis in

EkeGusiihelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 206

Figure 40 EkeGusii dissimilation from |γ| rarr [k]helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 218

Figure 41 Substitution of the English k for EkeGusii [g]helliphelliphelliphelliphelliphelliphelliphelliphelliphellip 220

Figure 42 Substitution of the English d for EkeGusii [t] English k for EkeGusii [g]helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 220

Figure 43 Realization of EkeGusii γ as [g] helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 226Figure 44 Realization of EkeGusii as [b] helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 227Figure 45 Realization of EkeGusii r as [d]helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip 227

Figure 46 English and EkeGusii treatment of nasal + consonant sequenceshelliphelliphellip 237

xxviii

CHAPTER ONE

INTRODUCTION

10 Background to the study

The purpose of this study is to examine the nature of the phonological and morphological

processes involved in the nativization of nouns borrowed from English into EkeGusii language

in order to understand the internal patterns of the phonology and morphology of these nouns The

focus of the study is on two languages in contact- EkeGusii the borrowing language and

English the lending language In this sub-section a brief background to the study is given

The sub-section begins with a brief description of the languages under focus in which an outline

of the languages in terms of where they are spoken by who and the language families to which

they belong Further it is shown how and why the two languages came into contact among other

highlights This is then followed the background of the study

According to Cammenga (200220) EkeGusii is ldquoan Eastern Nyanza Bantu language spoken in

the northern part of the region between the eastern shore of Lake Victoria or Lake Nyanza and

the eastern branch of the Great Rift Valley This region which includes the islands along the

lakeshore will be designated in geographical terms as East Nyanza so that EkeGusii may be

classified more specifically as an East Nyanza Bantu language It is labeled E 42 by Guthrie

(1971 vol 345 vol 311-15)rdquo EkeGusii language is spoken exclusively in Kenya However in

some literature including Cammenga it is wrongly observed that the language is more

commonly known by its English name viz Kisii (Ongarora 2009) The correct position is that

the language does not have an English name as claimed by Cammenga What Cammenga calls

the English name of the language (Kisii) is a name commonly used to refer to the administrative

and business town of the region or sometimes to refer to the entire region occupied by the

AbaGusii people Therefore as Cammenga points out ldquoin the language itself its proper name is

[ekeγusii] (EkeGusii) it is spoken by the [aβaγusii] (AbaGusii) peoplerdquo (2002 20) This is

supported by the observation by Demuth (2000) and Ongarora (2009) that names of the

languages in Bantu fall under classes 78 of the Bantu classification eke- therefore is a prefix

marking language Just like in Ongarora (2000) to be in conformity with the native speakers of

the language the name EkeGusii which bears the class prefix to denote language is used in this

EkeGusii boarders the following languages Kipsigis a Nilotic language to the north and partly

to the East Maasai Cushitic to the east and partly to the southeast DhoLuo which is Nilotic is

spoken to the west and southwest of EkeGusii and finally Kuria a Bantu language (E43) is

spoken to the south of EkeGusii and ldquoof all the East Nyanza Bantu Languages it is probably most

closely related to EkeGusiirdquo (Cammenga 200221) The language is spoken by over 2 million

AbaGusii (Kenya National Bureau of Statistics 2009) scattered world over

According to Bosire and Machogu (2013) EkeGusii is closely related to other Bantu languages

of the region described above and share common ancestry with KiKuria and Suba spoken in

Tanzania as well KiNyarwanda spoken in Rwanda and KiMeru KiKuyu and KiMaragoli all

spoken in Kenya

According to Crystal (1987) and Grimes (1996) English belongs to the Western Germanic

branch of the Indo-European group of languages Indo-European is the family of languages that

first spread throughout Europe and many parts of southern Asia and are now found in every part

of the world because of colonialism among other factors (Crystal 1987) This family Grimes

(1996) observe includes several branches Indo-Iranian Gothic Italic Baltic and Germanic

among others Of these branches the Italic and the Germanic are of significance as far as the

study of the development of English is concerned

According to Crystal (1987) English is a world language because of the political and socio-

economic progress the English speaking countries have made over the past 200 years Crystal

further observes that conservative estimates put mother-tongue speakers of English as 300

million second language users as another 300 million and fluent users as a foreign language as

100 million people Since this observation was made almost two decades ago users of the

language in the given areas could have increased a great deal by now According to Ethnologue

English is spoken by 983 522 920 people across the globe Thus English is indeed a world

language

Owino (2003) observes that English first came to Kenya through the Christian missionaries

traders and the British colonialists Currently it is estimated that there are approximately 900

million speakers of the language all over the world many of them being second language

speakers Ethnologue In Kenya like in most other African countries the language has few first

language users Majority of the users in the country are second language speakers who acquire

the language through the education system (Owino 2003)

Several definitions of the term nativization have been advanced by linguists For example

Haugen (1970432) sees nativization as the ldquoattempted reproduction in one language patterns

previously found in anotherrdquo Hock (1991390) defines it as ldquothe integration of foreign words

into onersquos native structuresrdquo while Mberia (2004) observes that nativization or borrowing is the

adjustments that borrowed words undergo at the various linguistic levels in order to be

accommodated in another language Definitions of nativization indicate that borrowed words

undergo certain linguistic processes in order to be accommodated or accepted in the target

language In other words the words undergo adaptation processes to conform to the structural

constraints of the borrowing language

For borrowing of words to occur languages must come into contact In other words the speakers

of the borrowing language must be in contact with those of the loaning language (Haugen 1970)

EkeGusii language has had such contact with a number of languages including Dholuo Maasai

and Kalenjin because of sharing a common ethnic border and English and Kiswahili due to

factors such as trade education and colonization among others As a result several nouns from

these languages have found their way into EkeGusii language as illustrated by (1) 1) EkeGusii borrowed nounsEkegusii borrowed noun source noun source language glosseγetaβu Kitabu Swahili book emesa meza Kiswahili tableesukuru School English school| eretio| eretjo radio English radio|riraβuani| rirawani rabuon Dholuo sweet potato ɳasae nyasaye Dholuo God Source Bosire amp Machogu (2013)These borrowed nouns come into the target language with different linguistic structures

(phonologically and morphologically) which means that for the loanwords to be accepted into

EkeGusii grammar they have to undergo readjustments structurally in terms of the various

linguistic processes This study deals with the phonological and morphological readjustment

processes that characterize nativization of EkeGusii loans from English

Loan word nativization processes have been studied extensively in among others Volland

(1986) Hock (1991) Yip (1993) Davidson amp Noyer (1997) Jacobs amp Gussenhoven (2000)

Kang (2003) Mberia (2004) Rose amp Demuth (2006) Mwita (2009) Owino (2003) Meiseburg

(1993) Hall amp Hamann (2003) Heinemann (2003) Dupoux amp Peperkamp (2002) Lacharite amp

Paradis (2005) Zivenga (2009) Raiz (2011) and Chebanne amp Phili (2015) However even if

this is the case the exact nature especially that of the internal phonology and morphology of

loan words in the target language is not well understood The questions being which

phonological and morphological language constraints do these loans obey (targetrsquos or sourcersquos)

given that each natural language has language particular ranking of the universal constraints

(Prince and Smolensky 1993) Which phonological and morphological language processes do

the loans words undergo in order to be accommodated in their new environments given that

some phonological and morphological processes are language specific affecting one language

and not the other and vice versa These are some of the questions that guided this study

11 Statement of the problem

The well-formedness of natural language is constraint-governed Even loaned lexical items obey

these constraints Constraints are universal and are ranked on a language specific basis thus no

two languages have a similar ranking of constraints The focus of this study is to investigate the

phonological and morphological adjustments that English nouns borrowed into EkeGusii

undergo to harmonize themselves with EkeGusii phonological and morphological systems

12 Objectives of the research

The purpose of this study is to contribute towards the understanding of nativization of the

phonological and morphological structures of English nouns borrowed into EkeGusii The study

seeks to achieve the following objectives

i) To describe the phonological and morphological structures of English and EkeGusii

ii) To analyse and discuss the phonological changes that English nouns borrowed into

EkeGusii undergo

iii) To analyse and discus the morphological changes that English nouns borrowed into

EkeGusii undergo

13 Research questions

The following questions will guide and aid the achievement of the objectives listed above

i) How dissimilar are the phonological and morphological structures of EkeGusii and

English nouns

ii) How does the EkeGusii phonological structure impact on the English nouns borrowed

into EkeGusii

iii) How does the EkeGusii morphological structure impact on the English nouns borrowed

into EkeGusii

14 Justification and significance of the study

Literature reviewed show that not much research has been conducted in EkeGusii in particular

in phonological and morphological nativization of loan words Existing literature further show

that research done on the language has focused on phonology morphology and syntax

(Whiteley1960 and 1965 Osinde 1988 Ongarora 1996 and 2009 Mecha 2006 and 2013 and

Nash 2011) among others No known study therefore has focused on nativization of EkeGusii

borrowed nouns from English or any other language This study therefore is pioneering in the

study of borrowed words into EkeGusii and it is significant because it acts as a safeguard to the

language which is exposed to an influx of loans (Zivenge 2009) Analysis of nativization of

English nouns borrowed into EkeGusii is significant because it focuses on how EkeGusii

modifies the English nounss in order to conform to its (EkeGusii) phonological and

morphological systems and status in a global linguistic enclave thus protecting the language

(EkeGusii) from extinction

Available literature indicates that many studies on nativization processes have been conducted

within the framework of Optimality Theory - a fairly recent constraint-based approach to

phonology However no such research has been conducted on morphological and phonological

nativization of English nouns borrowed into EkeGusii This study sheds light on the applicability

of Optimality Theory in accounting for the phenomenon of loan words nativization in EkeGusii

besides providing information to linguists who have interest of conducting constraint based

analyses on not only related languages but others as well

Many phonological and morphological theories would have been employed in such a study

However Optimality Theory a constraint based approach is used This theory is apt in this study

because its use of output (markedness) constraints such as CODA COMPLEX C which could

motivate the adaptation processes even when particular processes themselves have no

precedents in the native phonology (Yip 1993 Pardis amp LaCharite 1997 Jacobs and

Gussenhoven 2000 and Broselow 1999) This is because nativization involves adaptation of

foreign processes some of which not present in both the source and target language and

therefore could not have equivalent rules if rule based approaches are used This is one of the

strongest justifications for the use of the constraint based approach over rule based approaches in

this study

There are many studies targeting language borrowing generally and loan word nativization in

particular because language contact and the resultant lexical borrowing is an everyday activity

(Appel and Muysken 1987) This study does not only shed light into the phenomenon of

nativization but also adds knowledge to the growing and continuing investigations into the

Further this study is significant in that it is expected to shed more light on the phonological and

morphological structures of EkeGusii in particular and Bantu languages in general This is

because as (Owino 2003) observes the modification of the new words into EkeGusii makes it

possible to observe the nature of EkeGusii linguistic constraints which characterizes the

phonological and morphological structures of the language Hyman (1970) and Owino (2003)

suggest that lexical items of a given language do not provide as much opportunity in the study of

its grammar as much as new or borrowed items because the new items are necessarily modified

to fit into the constraints of the target language Observations into these adjustments reveal

structural features of the target language ( EkeGusii in this case)

Another significance of this study is that its findings will shed light on the structure of other

Bantu languages that have similar formal content as EkeGusii thus enabling linguists to

understand the phonology and morphology of those languages and widen their understanding of

nativization and the processes involved

15 Scope and limitation

This study examines English nouns borrowed into EkeGusii and their phonological and

morphological adjustments in the process of their accommodation into the phonological and

morphological structures of EkeGusii language An attempt is made to define and rank the

constraints determining phonemic syllabic (phonotactic) phonological processes and tonemic

adaptation of English phonological features entering EkeGusii Further borrowed words also

undergo morphological changes upon entering the target language This study focuses on those

morphological features that affect the noun class only including inflections (affixations) for

number and person as well as classification of the nouns into various word classes present in the

recipient language

Semantic changes also characterize loan words (Bloomfield 1933) However this particular

aspect of loan word adaptation is not examined Thus this study does not focus on semantic

adjustments on the borrowed nouns but instead the role of semantics in as far as it characterizes

morphological nativization is examined Only phonological and morphological adjustments are

studied as they may offer interesting clues to the understanding of the phonology and

morphology of borrowed words (Hall and Hamann 2003)

EkeGusii language has had contact with many languages such as Dholuo Kalenjin English and

Kiswahili among others However this study focuses only on borrowed nouns from English

language because as Scotton and Okeju (1972) and Owino (2003) observe the language of

numerically and socio- economically and culturally more dominant peoples are the more likely

donors in lexical borrowing while the less prominent groups are more often the borrowers This

is the case in the Kenyan situation where EkeGusii although surrounded by other numerically

dominant groups like Kalenjin it does not borrow from them as much as it does from English

This is because unlike English these languages are less dominant socio-economically and

prestige (Owino 2003) Although according to Owino there are a few first language speakers of

English in Kenya the language is prestigious and occupies a culturally and socio-economically

dominant position and therefore an obvious donor for most Kenyan indigenous languages

including EkeGusii This observation coupled with the fact that English is an official language

in the country makes English the choice for this study

The study focuss on the processes of phonological and morphological nativization of English

nouns in EkeGusii There are many of such processes such as phoneme change resyllabification

among others in phonology and noun formation affixation among others in morphology

However only those phonological and morphological processes that directly affect nativization

like affixation and resyllabification are described and analyzed given the nature of this study

Focusing only on the relevant phonological and morphological processes will give ample time

and space to focus on concepts that give insight into nativization processes for purposes of detail

and precision At the phonological level the study will focus on the constraints that determine

re-syllabification phonemic change and phonological processes such as vowel harmony and

Dahlrsquos law that characterize English nouns in EkeGusii

Morphologically several processes characterize nativization of lexical items in the borrowing

language Such processes include affixation involving tense mood and pluralization among

others across the main parts of speech However this study does not focus on all these features

in all the parts of speech instead it is confined to some affixation processes relating to the noun

class only because nouns are the most loaned of all the word classes (Hockett 1958) The

processes on focus are prefixation augmentation and nominal classification

EkeGusii language has two dialects EkeMaate and EkeRogoro dialects (Mecha 2013 Ongarora

1996 and Bosire 1993) This study is confined to the EkeRogoro dialect of EkeGusii because it

has the majority of speakers of the language

Lexical items the units of focus in this study vary widely in terms of domains of life food and

nutrition house-hold appliances and utensils clothing transportmotoring information

technology education politics and governance health diplomacy business and trade legal

affairs agriculturefarming and security and administration The study targeted English nouns

borrowed into EkeGusii from at least all the identified domains of life without exception

CHAPTER TWO

LITERATURE REVIEW AND THEORETICAL FRAMEWORK

20 Introduction

This study focuses on linguistic borrowing which has been defined variously by different

language contact scholars For example Crystal (1987) sees borrowing as the introduction of

words and other linguistic features by one language to another To Hoffer (2005) borrowing is

the process where linguistic items of one language are transferred into another language whereas

Thomason and Kaufman (2008) view borrowing as the transferring of linguistic materials

(lexical phonological morphological semantic and syntactic) from a donor language to a

recipient language as a result of contact between languages Lexical borrowing therefore deals

with the transfer of words across language boundaries (Bynon 1977) It is these lexical

borrowings that are referred to as loan words (Kang 2011)

Peperkamp and Dupoux (2001) identifies two types of loanwords Historical loanwords that is

those that have entered the borrowing language and are commonly used by monolingual speakers

who never hear source forms of the loan words The other type of loan words is the online

adaptation or on-going adaptation This according to Peperkamp eventually becomes historical

when fully adopted into the borrowing language This study focuses on historical loan words in

which monolingual speakers of EkeGusii were the source

Another category of the types of borrowing has been advanced in the literature Here there are

types such as direct borrowing in which the borrowing language adopts and integrates into its

system both the form and meaning of a lexical item from another language (Bynon 1977 Asher

1994 and Owino 2003) Bynon (1977) identifies loan translation or calques as another type

Here the meaning of a foreign lexical item is employed as a model for a native creation instead

of being carried over (Owino 2003) Other types in this category include semantic extensions

loan shift and loan blend among others (Bynon 1997 Asher 1994)

A number of reasons have been advanced to explain why languages borrow from one another

For example Hockett (1958) Weinreich (1963) and Owino (2003) give the prestige motive as

one of this motivations This is a case where the borrowers emulate the people they admire by

borrowing linguistic items from them Other reasons advanced include the need to fill motive

(Langacker 1968) the need to designate new things (Weinreich 1963 Taber 1979 Owino

2003) and the need to resolve a clash of homonyms (Weinreich 1963) These reasons apply in

this study

According to Danesi (1985) nativization is the entire process of the adaptation of a loan word by

a speaker Danesi observes that loan words are adjusted unconsciously and systematically to the

pronunciation and grammatical patterns of the borrowing language He further points out that

loaned words are not accepted in their original shapes but rather are restructured to conform to

the articulatory and grammatical features of the borrowing language

Studies of this nature according to Owino (2003) are important in as far as understanding of the

internal structure (in terms of phonology morphology syntax and semantics) of loan words is

concerned Thus borrowing processes have often been studied with the hope that they may

reveal internal structural features of the borrowing language This is because loan words usually

undergo adaptation processes to conform to the structural constraints of the borrowing language

phonology

This chapter reviews some of the studies conducted in the area of linguistic borrowing so far

with the aim of establishing and filling in existing knowledge gaps The review contextualizes

this study and brings forth better understanding of the debates controversies and current issues

in the area of phonological and morphological nativization The chapter reviews literature related

to the study and the theoretical framework employed It reviews literature on EkeGusii

phonological and morphological structures in order to shed light on the phonology and

morphology of the language in order to show how they differ from those English- the loaning

language Literature on lexical borrowing focusing on nativization related studies on Bantu and

non Bantu languages conducted within and outside OT are reviewed with the aim of establishing

existing gaps that need to be filled The chapter is subdivided into the following subsections 21

which addresses objective one of the study 22 deals with objective two 23 objective three

and 24 which deals with the theoretical framework within which the study was conducted

21 EkeGusii phonological and morphological structures

In this subsection literature related to objective one of the study is reviewed The objective aims

at describing the phonological and morphological structures of EkeGusii which sets the basis of

the analyses in subsequent subsections The sub-section is divided into 211 phonological

structure of EkeGusii and 212 morphological structure of EkeGusii

211 EkeGusii phonological structure

This subsection reviews literature on the phonological structure of EkeGusii It focuses on the

phoneme structure the syllable structure and the toneme structure Studies conducted in

EkeGusii phonological system so far (Whiteley 1960 Cammenga 2002 amp Ongarora 1996)

among others show that the language has seven vowels This study is in agreement with these

studies that EkeGusii has seven vowels that is a e ε i o ͻ u However unlike the pioneering

studies above which give mere descriptions of the vowels this study gives a spectrographic

analyses of the vowels using the Praat software (Boersma ampWeenink 2010) These analyses give

the exact acoustic nature of the vowels unlike the early impressionistic analyses of the vowels For

example the exact height and mouth positions of the vowels as recorded from native speakers are

determined This indeed is a milestone in as far as the study of EkeGusii phonetics and phonology

is concerned

EkeGusii is characterized by height harmony (Ongarora 1996) Height harmony according to

Ongarora and in many other studies is based on the impressionistic vowel triangle (Keith 1997)

However in this study the vowels perceived to be in harmony are spectrographically analyzed and

determined using spectrograms This is based on the fact that vowel height is negatively correlated

with formant 1 (F1) frequency [ High] vowels have low F1 and [low] vowels have high F1

(Keith1997105)

Phonotactics of a language are the constraints on the sequence or position of phonemes in words in

that language Such constraints are part of every speakerrsquos phonological knowledge of their

language The constraints operate on units larger than the single segment or phoneme that is the

syllable (Yule 1996) Consonant phonemes generally operate at the margins of syllables either

singly or in clusters (Mohamed 2000) in languages In EkeGusii just like in most Bantu languages

(Mutua 2007) all consonants function as onsets and not codas Thus as Wald (1989) and Mutua

(2007) observe prenasalised series common in Bantu should be treated phonologically as an

independent series rather than as a cluster of nasal plus stop Therefore EkeGusii [mb] [nd] [ng]

and [nt] are independent (single) phonemes and not consonant clusters (Herby 1986 amp Downing

Pioneering studies in EkeGusii phonology (Whiteley 1965 amp Cammenga 2002) among others

show that EkeGusii has consonant glide sequences or consonant clusters in short This study

however following observations by Herby (1986) Hayman and Katamba (1999) Order (1999)

and Downing (2005) suggests that EkeGusii like other Bantu languages lack consonant clusters

Rather what seems like consonant glide sequences are derived secondary consonants Thus where

for example there is a consonant glide [CW] sequence in Cammenga (2002) this study sees it as a

derived secondary articulation [Cw] Studies in other Bantu languages support this view For

example in LuGanda spoken in Uganda the first vowel in a word is deleted unless it is high (in

which case it becomes a glide [w] or [j] (Katamba 1993) The same is true of Ronga (Hargus and

da Conceicao 1999) Emai (McCarthy 2007) and Fahiru (Otterloo 2011) This observation is

important to this study because it explains why borrowed words with consonant clusters from

English are declustered and the extent of declusterization The observation further emphasizes the

fact that EkeGusii like other Bantu languages is a strict CV language Thus any form of consonant

clusters is not allowed

A number of consonantal processes affecting EkeGusii language and Bantu languages generally

have been described in a number of studies and scholary papers For example Uffmann (2013)

makes a general observation that voice dissimilation (Dahlrsquos law) affects a number of Bantu

languages Guthrie (1967) outlines some of the languages affected in which EkeGusii is included

This is a process in which consonants in a given word become dissimilar in terms of voice

depending on the language in question Other consonantal processes which affect Bantu languages

besides voice dissimilation include Declusterization in which consonant clusters are broken

(Hyman and Katamba 1999 Oden and Oder 1999 McCarthy 2007 and Otterloo 2011)

defricativization in which a fricative consonant becomes non fricative by losing its continuant

characteristics (Cammenga 2002) and nasal resyllabification in which nasals optionally become

syllabic and bear tone (Ferguson 1963amp Hyman 1985) These consonantal processes are

important to this study because they have guided in establishing the extent to which Bantu

phonologies in general and EkeGusii phonology in particular affect borrowed consonants

especially those from English

Pioneering studies in EkeGusii phonology (Whiteley 1965 Ongarora 1996 and Cammenga

2002) show that EkeGusii like most other Bantu languages such as Shona (Kadenge 2003)

Kiswahili (Mwita 2009) and Tonga (Zivenge 2009) among others is a strict CV language It

does not allow consonant clusters English on the other hand is not a strict CV language because it

sometimes allows consonant clusters of up to three and sometimes four consonants as in the words

structure srkatintǝ and sixths sikss respectively (Roach 1983 and Cruttenden 2011) This

observation is important to this study because it allowed observation into how consonant clusters

from English which are not allowed in EkeGusii as in other Bantu languages are dealt with by

EkeGusii

Tone according to de Lacy (2007) is the use of pitch in language to distinguish lexical or

grammatical meaning While some languages are characterized by this feature and therefore

described as tone languages others are not Thus in tone languages pitch differences are used to

make lexical meaning differences This is unlike in non-tone languages which are characterized by

stress This is an important observation to this study because EkeGusii which loans words from

English is a tonal (Cammenga 2002 Bickmore 2007 and Nash 2011) while English is a stress

language (OrsquoConner 1967 Roach 1983 Oden 2005 Cruttenden 2011 and Gussenhoven and

Jacobs 2013) Thus it has allowed observation into how EkeGusii loan words from English a

stressed language are handled by EkeGusii a tone language

212 EkeGusii morphological structure

According to Meinhoff (1899) nouns in Bantu languages are morphologically classified into sets

of classes Bantuists have given varied numbers of such classes ranging from 22 in pro-Bantu

(Welmers 1973) to as few as 12 in Aghem spoken in Cameroon (Aikhenvald 2000) The

observation that nouns in Bantu are grouped into sets of classes is important to this study because it

allows observation of how EkeGusii loan words from English in which nouns are not grouped into

classes are handled in EkeGusii

Morphosyntactically Bantu nouns consist of a prefix and a stem which are normally compulsory

The prefix according to Cammenga (2002) contains number and size features The choice of these

prefixes is determined by the semantics of the noun stems with which they occur (Cammenga

2002 Kayigema 2010) This is an important observation to this study because it shows how

EkeGusii loan words from English in which prefixation neither marks size and number features

nor determined by noun stems are dealt with

Elwell (2005) observes that structurally most noun prefixes in EkeGusii like in some other Bantu

languages Kinyarwanda (Kayigema 2010) and Kuria (Mwita 2009) are divided into two parts

a pre-prefix (augment) and a prefix This observation about the morphological structure of the

Bantu nouns generally and those of EkeGusii specifically is equally important to this study This is

because it has allowed observations into how EkeGusii loaned nouns from English with different

morphological structures are handled by EkeGusii

22 Phonological nativization

This sub-section rviews literature related to objective two of the study which analyses the

phonological changes that EkeGusii loan words from English undergo during nativization

Phonological nativization involves the phonological adjustments that a loan word undergoes in

the borrowing language in order to be accommodated in the phonological structure of that

language

Many studies have been conducted in the area of phonological nativization For example Hock

(1991) discusses phonological nativization of loanwords in general terms without making

reference to any specific languages In the study Hock discusses the phonological processes that

characterize nativization generally He observes that borrowed words have to change

phonologically to suit the phonological characteristics of the receiving language He emphasizes

on the substitution of the various corresponding loaner language phonemes for those of the

receiving language However Hockrsquos study unlike the present one is not grounded on any

theory It also deals with phonological nativization only while the present one deals with

morphology as well These are the departing points between the two studies Hockrsquos study will

give insight in as far as possible phoneme changes are concerned in this study

Hall and Hamann (2003) discuss phonological loan word nativization in German in which they

focus on changes that affect stressed vowels in the nativization of loanwords from Italian

language into standard German Their conclusion is that the Italian tense or long vowels such

as i and ͻ are laxed or shortened in the process of nativization especially when they occur

between voiced consonants This study gives a lot of insight into the current one because firstly

it deals with phonemic change in the nativization processes a process that affects the current

study and secondly both studies are anchored on a theory- Optimality Theory However

whereas Hallrsquos amp Hamannrsquos study deal with vowel phonemes only the present study deals with

consonants as well Hall amp Hamann (2003) discuss phonological nativization only while the

present one deals with morphological nativization as well

Miao (2005) studies Phonological loan word adaptation in Mandarin Chinese within the

perspective of OT He focuses on phoneme substitution patterns for consonants and re-

syllabification processes in Mandarin Chinese loans from three languages- English German and

Italian Miaorsquos study is similar to this study in that both studies deal with phoneme change and

re-syllabification processes Both studies are anchored on OT However the studies differ in a

number of ways For example the target languages are different in Miaorsquos it is Mandarin

Chinese while it is EkeGusii in the current study Another point of departure is that Miao

considers only consonant phonemes while this study considers vowels as well Finally while

this study investigates the possibility of the influence of Dahlrsquos Law in phonemic change during

nativization Miarsquos does not

Uffmann (1993) discusses vowel epenthesis in Shona loanwords from English In his analyses

he shows that vowels are epenthesized on an acceptable consonant sequences (consonant

clusters) According to Uffmann Languages with strict CV syllable structure often epenthesize

vowels in positions where they serve to break up consonant clusters or re-syllabifying coda

consonants as shown in (2) adapted from Uffimann

2) Vowel epenthesis in English nouns in Shona

a) Coronal + i [bazi] lsquobusrsquo

b) Labial + u [temu] lsquotermrsquo

In these nouns vowels have been epenthesized ([i] in lsquobuzirsquo and [u] in lsquotemursquo) in order to avoid

codas which are allowed in English but not in Shona because Shona unlike English is a CV

language Shona like EkeGusii (which is the focus of the current study) is a Bantu language

Thus the current study gets insight from the Shona study since both studies deal with phonology

The departing points between these studies include while Uffimann (1993) deals with vowel

epenthesis only the current study deals with other processes as well phonemic change re-

syllabification among others Finally the current study unlike that of Uffimann which deals

with phonology only deals with morphological nativization as well

Kim (2008) gives a phonological highlight of Huave language loans from Spanish Among the

conclusions he draws are that Spanish the loaning language of Huave has an open syllable

structure while Huave has a closed syllable structure This prompts apocope (ie final vowel

deletion to Spanish loans in Huave The exact opposite characterizes English loans in EkeGusii

This is because English the loaning language of EkeGusii is sometimes a closed syllable system

while EkeGusii is a purely open syllable system Besides apocope this study describes other

phonological processes Syllable deletion which according to Kim is a situation where an entire

final syllable is deleted especially in words where Spanish stress as antepenultimate

palatalization- this takes place in the environment of tautosyllabic front vowels Other processes

include cluster resolution fortition (where intervocalic voiced fricatives become voiceless stops)

vowel leveling (a case where one vowel overrides and changes the quality of a neighbouring

vowel which becomes a copy of the trigger vowel -vowel harmony)

Discussing morphological loan word nativization Kim (2008) observes that generally nouns are

borrowed as bare roots from Spanish into Huave with only phonological modifications Kim

gives more attention to verb morphological nativization

Kimrsquos study gives the present study a lot of insight even though his major focus is not

nativization per-se Phonological processes studied in both studies are the same to a large extent

However while Kim focuses on verb morphological adaptation this study focuses on noun

morphological nativization as well

The major departing points between these two studies is that Kimrsquos study unlike the present one

is not anchored on any theory Kim merely describes the various phonological and

morphological processes without any theoretical grounding Descriptions and analyses in the

present study are based on Optimality theory

Zivenge (2009) studies phonological nativization of Tonga language Like in the current study

Zivengersquos study is anchored on theories unlike many others reviewed already However whereas

at the phonological level Zivenge grounds his study on the rule based CV phonology theory the

present study is grounded on the constraint based Optimality Theory Another difference

between the two studies is based on phoneme change The present study unlike that of Zivenge

investigates the influence of the dissimilatory process (Dahlrsquos Law) in EkeGusii loan word

nativization process This is because as Ellwel (2005) observes EkeGusii unlike Tonga is

among the few Bantu languages characterized by this dissimilatory process Finally while

Zivange (2009) deals with Tonga a Zimbabwean Bantu language this study deals with

EkeGusii a Bantu language spoken in Kenya

Another study that benefited this study immensely is that of Mwihaki (1998) This study deals

with Gikuyu loaned words from English and identifies three aspects of loaned word adaptation

phonemic phonotactic and prosodic The current study focuses on phonemic and phonotactic

changes that characterize EkeGusii loaned words from English Like Mwihakirsquos this study

considers the changes that loaned words undergo at the syllable level However unlike

Mwihakirsquos the current study does not assign prosodic features to the loans The major departing

point between these two studies however is in their theoretical orientation While Mwihakirsquos is

grounded on Auto Segmental and Metrical Phonology this study employs the constraint based

Optimality Theory in its presentation and analysis of data Again while Mwihaki studies only

phonological adaptation the present study deals with morphological adaptation as well

Mberia (2004) discusses the phonological behavior of borrowed words in Kitharaka He focuses

on Kiswahili and English loans This study differs from the current one in the following ways

while Mberia merely discusses the phonological processes involved during nativization without

any theoretical grounding the present study is anchored on a theoretical framework Mberiarsquos

study however gives a lot of insight into this study because both Kitharaka and EkeGusii are

Bantu languages (Guthrie 1967-71) thus the phonological and morphological behavior of the

English loans into both languages show similar trend Besides Mberia shows that Kitharaka like

EkeGusii is affected by Dahlrsquos law and therefore the study is instrumental in understanding the

operation of Dahlrsquos law in Bantu languages

Owino (2003) deals with phonological nativization of Dholuo loanwords This study in as much

as it gives insight into the general topic of nativization differs from the current study in

significant ways Firstly the two studies deal with languages from different groups Dholuo is a

Nilotic language while Ekegusii is a Bantu one (Guthrie 1967-71) This means that the English

loans might have quite different characteristics in Dholuo as compared to EkeGusii since the

phonotactics of the two target languages differ quite significantly Secondly the current study

unlike Owinorsquos deal with two linguistic areas ndash morphology and phonology Finally while the

proposed study will be anchored on the constraint based OT paradigm Owinorsquos is anchored on

Auto segmental and CV ndash Phonology theories

Mutua (2007) analyzes Kikamba nativized loanwords in which he analyses Kikamba English

loanwords This study is insightful to the present study in that both studies deal with Bantu

languages and are both anchored on OT However the two studies differ in that whereas Mutuarsquos

study deals with only phonological nativization the present study deals with morphological

nativization as well Secondly the present study studies voice dissimilation in EkeGusii

loanwords and the role of noun classes in the nativization process Mutuarsquos does not

The present study like those ones of Zivenge (2009) Lodhi (2000) Chimhundu (2002) Mwita

(2009) among others deals with a Bantu language Zivenge studies Tonga language Lodhi and

Mwita Kiswahili while Chimhundu researches on Shona language This means that the

characteristics of English loans in these languages are insightful in predicting the characteristics

of EkeGusii loans from English This is because the languages in the given studies like EkeGusii

are Bantu and therefore have a lot in common in terms of linguistic properties (Shillington

1995) But this does not mean that the languages are similar since they have what are referred to

as unique values (Bloomfield 1933)

Bickmore (1997 and 1999) deal with EkeGusii verb tone within Optimality Theory while

Cammenga (2002) describes EkeGusii verb tone without using any theoretical grounding Nash

(2013) analyses EkeGusii noun and verb tonology Like Cammenga Nashrsquos analyses are not

anchored on any theory The present study like that of Bickmore is anchored on OT But while

Bickmorersquos analyses the verb this study analyses the noun

23 Morphological nativization

This subsection reviews literature related to objective three of the study which analyses the

morphological adjustments that English loan words into EkeGusii undergo in order for them to

be accommodated into the morphological structure of the language Morphological nativization

involves the morphological changes that borrowed words undergo in order to be accommodated

in the morphological structure of the target language Islam (2011) studies the morphology of

Urdu loanwords from English Arabic and Persian He concludes that the affixation of English

loans into Urdu whether inflectional or derivational tend to be on native Urdu patterns Further

that English loans are adaptable and that their integration takes place with native Urdu affixes

which are productive and conveniently attach to affixes The study established that both

inflectional and derivational changes are involved in the loaning process This study is similar to

the present one in a number of ways they both deal with morphological processes affecting

loaned words Like in Islamrsquos study where affixation tends to be on Urdu patterns affixation in

this study also follows EkeGusii patterns However the two studies differ as follows While the

current study is anchored on a theory (OT) Islamrsquos study is just descriptive it is not anchored in

any theory Secondly while in Islamrsquos study there are both derivational and inflectional affixes

in the nominal class nouns affixations in the current study is limited to only inflectional affixes

Thus derivational affixes do not characterize the loaned words in the current study This is

because unlike Urdu which allows suffixation EkeGusii does not Thirdly while the prefix in

EkeGusii is characterized with a pre- prefix the same is not a feature in Urdu because pre-

prefixing is a feature present in a few Bantu languages

Zivenge (2009) discusses the morphological features which characterize Tonga loans from

English in the perspective of Lexical Phonology and Morphology Theory The current study on

the other hand studies morphological features that characterize EkeGusii loans within the

perspective of OT In nominal nativization Zivenge focuses on the class marking prefix which is

characteristic of most Bantu Languages (Elwell 2005 Bresnan and Mchombo 1987) This gives

insight into the present study since EkeGusii like Tonga is Bantu and is characterized by

nominal class prefixes But the two studies differ in a significant way concerning nominal class

prefixing EkeGusii noun class prefixing is characterized by an augment or pre-prefix unlike

Tonga Thus EkeGusii nominal loans from English are not only prefixed but also pre-prefixed

as well to mark class number and size This is another departing point between the two studies Kayigema (2010) analyzed how French and English loaned words into KiNyarwanda are

allocated to what he calls key areas of influence and nominal class systems of KiNyarwanda

The study established that bilinguals play and important role in importing of foreign words to the

target language that loaned words from languages closer to the language system of the

borrowing language adopt more easily than those from more distant languages The study

confirmed that some of the key areas into which loaned words are allocated include commerce

technology transport agriculture among others It further established the most affected parts of

speech during borrowing are nouns and verbs because at least all languages of the world have

them besides they express more concrete realities as compared to the other part of speech

These findings are beneficial to the present study because they help in the focusing of the areas

of influence identified

Regarding nominal class allocation of loaned words Kayigema (2010) observes that all the

foreign nouns entering KiNyarwanda enter into the nominal classes of KiNyarwanda thereby

adapting its morphological system For example nouns within the semantic field of [+ human

+singular] enters classes 1 and 2 as in u-mu-avoka and a-ba-voka for French avocat (lawyer) and

avocats (lawyers) respectively This observation is beneficial to the present study because

EkeGusii like KiNyarwanda is a Bantu language characterized by nominal classes Thus The

present study gets insight from Kayigemarsquos in as far as nominal classification of loaned words is

concerned because the nominal classifications in the two languages are likely to share some

characteristics

In as much as Kayigema (2010) benefits this study the two studies are different in a number of

ways Firstly the languages under focus (KiNyarwanda and EkeGusii) are different though from

the same language family (Bantu) Secondly Kayigemarsquos unlike the present study focuses on

the areas of loaned words allocation that is the areas into which the borrowed loans enter and

the nominal classes to which they are borrowed The focus of the present study however is in the

morphological changes that the loaned words undergo besides allocations into the various

nominal classes one such change being nominal classification which does not characterize

English noun morphology Finally the present study unlike Kayigemarsquos analyzes data using a

theory Optimality Theory In other words the morphological changes realized in this study are

analyzed and explained theoretically unlike in Kayigemarsquos

24 Theoretical framework

This study is anchored on concepts and ideas drawn from Optimality Theory (OT) as expounded

by Prince amp Smolensky (19932004) and McCarthy (2006) in its analysis of data McCarthy

(2006) observes that OT is a general approach to modeling human linguistic knowledge The

central argument of this theory that benefits this study is that surface or output forms of language

reflect resolutions of conflict between competing constraints that a surface form of language is

optimal if it incurs the least serious violations of a set of constraints taking into account the

different hierarchical ranking of constraints by languages (Prince and Smolensky 1993)

241 Tenets of Optimality Theory

The Optimality Theory is different from the rule based generative theories of phonology in a

number of ways (Prince and Smolensky 19932004) For example OT and the Principles and

Parameters Theory (Chomsky 1981) differ in in that while the two theories view grammatical

principles as universal they elaborate the principles differently Parametric theory sees the

principles as a set of inviolable constraints while OT sees them as a set of hierarchically ranked

and violable constraints Thus while language typologies are obtainable through parameter

setting (switching onoff of a constraint) in Principles and Parameters Theory the same is

achievable through re-ranking of violable constraints in OT (Kager 1999) The present study

sees languages as differing in this sense that is in the re-ranking of universal constraints

OT dictates that an optimal output form is selected from a set of candidates based on a (re)-

ranking of violable well formedness constraints the candidate that minimallyleast violates the

constraints in the given ranking (which is language particular) is selected as the optimal

candidate and thus appears as the surface form These candidates are evaluated in parallel instead

of subject to a series of ordered rules ndash as in rule based theories Additionally the set of

constraints in OT is proposed to be universal and that the grammars of languages theoretically

differ in the ranking order of the constraints For instance a highly ranked constraint in one

language (for example CODA in EkeGusii) may be lowly ranked in another language (such as

English) According to the theory a violation of a highly ranked candidate is fatal which means

that such a candidate will never be optimal The opposite is true

According to this theory there are two main types of constraints Constraints on the form of the

output structure (the well formedness constraints) on segments and segment organization These

being constraints grounded in universal markedness principles such as syllables must have onsets

and constraints on the relationship between the input and the output aimed at the preservation of

information (maintaining faithfulness of the output to the input) (Kager1999) Kager observes

that these two constraints are inherently in conflict

This theory has three key components Generator (GEN) the component which takes an input

and generates a list of possible outputs called candidates (possible realizations of an input which

are potentially infinite in number)

Constraint (CON) is another component This provides the criteria in the form of strictly ordered

violable constraints used to decide between candidates These constraints are assumed to be

universal Universal in the sense that they affect all languages though each language ranks them

differently (which is one of the reasons behind language differences)

The third component is Evaluator (EVAL) This is the component that choosesselects

depending on the grammar (language in question) the optimal candidate Each candidate is

evaluated by all constraints at once in parallel rather than in a serial fashion of the derivational

generative frameworks The candidate (output) that violates the fewest high ranked constraints is

chosen as the optimal by the grammar Evaluation takes place by a set of hierarchically ranked

constraints in the form (C1 gtgt C2 gtgt hellipCn) each of which may eliminate some candidate output

until a point is reached at which only one output candidate survives This elimination process is

represented schematically in figure (1)

candidates constraints

Input GEN

Figure (1) Process of candidate elimination in OT

Source Kager (1999)

According to McCarthy (2006) OT has had significant impact on various fields of linguistics

including phonology and morphology This study employs in its presentation and analysis of

data the premises of the theory that are most directly applicable to phonology and morphology

Some of the constraints from the universal set reproduced in Table (1) were used to present and

analyze phonological data generated in the study

Table (1) Some OT Constraints from the universal set

CONSTARINT INTERPRETATIONCOMPLEX (C) No complex marginsNo consonant clusters

chellip

gtgt gtgt

Output

COMPLEX (V) No strings of vowelsNo complex vowelsFAITH C The consonants in the input are the same as the consonants in the outputFAITH V The vowels in the input are the same as the vowels in the outputIDENT-IO (F) The specification for the features of an input segment must be preserved

in its output correspondentINDENT-IO (place) The specification for place of articulation of an input segment must be

VOI Voiced obstuents are not allowed

VTV Voiceless obstruents are banned in intervocalic positions

MAX- IO Input segments must have output correspondents (no deletion)MAX- IO V Input vowels must have output correspondents(no deletion)CODA Syllables are open ( must not have codas)ONSET Syllables must have onsets

DEP No epenthesis or addition of either a vowel or consonant

OCP(dorstop) A markedness constraint which bans sequence of voices dorsal stop

OCP(V) A markedness constraint which bans sequence of voiced consonants

COPLEX(C) A constraint which bans complex consonants

COPLEX(V) A constraint which bans a string of vowels

BACK V Back vowels are not allowed

CENTRAL V Central vowel are not allowed

LAX V Lax vowels are not allowed

HIGH V High vowels are not allowed

AGREE V Vowels in a word should agree

AGREE C Consonants in a word should agree

STRIDENT (FRIC) Strident fricates are not allowed

INTERD (FRIC) Interdental fricatives are not allowed

LATERAL (C) Lateral consonants are not allowed

TRILL (C) Trill consonants are not allowed

VOICEOBS Voiced obstruents are not allowed

NONANTERCORFRIC Non-anterior coronal fricatives are not allowed

ASSIM (F) Assimilation of features are not allowed in a given domain

UNIFORMITY Prohibits feature fusing

FLOAT Markedness constraint which prohibits floating of tones

SPECIFY A constraint which demands that a tone bearing unit must have a

correspondent tone

Source Archangeli (1997) Kager (1999) Prince amp Smolensky (2004) McCarthy (20072008)

Morphologically Optmality Theory provides insight into various morphological phenomena

including affixation reduplication and allomorphy (McCarthy 2006) However this study

employs the tenets of OT that are most directly applicable to the morphology of loanword

nativization affixation alignment constraints constraint ranking and violability competition

among candidate outputs faithfulness and parallelism of evaluation McCarthy (200638)

observes ldquoconstraint violability is pervasive in applications of OT but there are two areas of

morphology in which it assumes particular importance affix location and Template

morphologyrdquo These will be the areas of focus in this study

According to McCarthy and Prince (1993) affix alignment constraints demand that the edge of

two constraints coincide In particular a constraint requiring that the left edge of an affix align

with the right edge of a word (ALIGN (Affix L Word L)) has the effect of declaring an affix a

suffix while a constraint requiring that the right edge of an affix align with the left edge of a root

(ALIGN (Affix Root L)) will have the effect of declaring this affix to be a prefix McCarthy

and prince further observe that constraints on affixal alignment have also been applied to clitic

and affix order restrictions for example align the right edge of an affix to the left edge for a

word (ALIGN (Afx R Root L)) This study analyzes affix location of EkeGusii loan words

from English vis-agrave-vis the stated alignment constraints

The input the supposed underlying form of a grammar plays a crucial role in this theory

According to Prince and Smolensky (1993) the input has two main functions to determine the

output candidates which compete for optimality and to be referred to by faithfulness constraints

that prohibit output candidates from deviating from specifications in the input

This study utilizes the correspondence framework of Optimality Theory by McCarthy and Prince

(1995) McCarthy (2008) This framework provides that both input and output consist entirely of

overt non-abstract phonological material It gives a relation between the input segments and the

output segments that is correspondence (input- output correspondence) This framework rejects

abstract outputs and strengthens the notion input ndash bringing on board input optimization

arguments of Prince and Smolensky (1993) Input or lexicon optimization framework provides

that an output is faithful to an input This observation is demonstrated by the change of Old

English sk to modern English int as in scip [skip] rarr ship [intip] (de Gruyler 2002) In OT and

particularly in the input optimization approach this change means that the input as well as the

output are the same ([intip]) Thus the faithfulness constraints such as MAX IO is obeyed at the

expense of the markedness constraintCOMPLEX C (MAX IO gtgt COMPLEX C) as analyzed

in tableau (1)

input intip MAX IO COMPLEX C

a) intip

b) skip

Tableau (21) Modern English realization of the input intip

This tableau shows that the input has been optimized that is it has been realized without any

change and therefore is faithful to the output It is therefore an input as well as an output This is

how this study treats EkeGusii loaned words from English

242 Optimality Theory versus other Generative Approaches

Many theories could be employed in such a study generative and non-generative However this

study employs the constraint based generative phonology and morphology approach Optimality

Theory (Prince and Smolensky 1993 and McCarthy 2006) as compared to rule based

generative theories

Rule based generative phonology theories according to Clark Yallop and Fletcher (2007)

embodies the derivational approach whereby the output or surface form is the result of a series

of rules that operate on an underlying form of a word or morpheme generating at each stage of

the derivation a specific output which is in turn operated on by the following rules in the

derivation process In constraint based approaches such as Optimality Theory on the other hand

a surface form is realized not through rule application but by violating the least of a set of

language specific hierarchically ranked constraints which are violable

Optimality Theory is apt in this study because its use of output (markedness) constraints such as

CODA COMPLEX C could motivate the adaptation process even when particular processes

themselves have no precedents in the native phonology (Yip 1993 Pardis amp LaCharite 1997

Jacobs and Gussenhoven 2000 and Broselow 1999)

According to Smolensky (1996) the principle of the richness of the base in OT naturally allows

for and perhaps even requires an analysis of novel input forms which are not attested in native

learning data This makes the theory more amenable to the study of loaned word adaptation

phenomenon because all the loaned words from English (which constitute inputs) into EkeGusii

are novel

The fact that OT allows for the formalization of tendencies can be seen as beneficial over rule

based theories because phonologists have for long argued that tendencies (such as the historical

tendency towards consonantal lenition) or for stress to fall on heavy syllables

Again generative phonology of the 1970s and 1980s had increasingly developed a mixed model

which used both rules and constraints OT unlike these generative phonological theories enables

phonological entity (constraints) This is an advantage according to Arbib (2006)

OTrsquos attempt to account for opacity such as in Sympathy Theory where failed candidates are

allowed to influence the successful candidate and Stratal OT which introduces lexical strata has

rescued much of what was proposed in the model of lexical phonology

OT unlike rule based theories predicts the emergency of the unmarked (TETU) Thus a

markedness constraint that is frequently violated in a language may still affect output According

to Arbib (2006) the constraint favoring CiCC over CCiC in the language of Yawelmani for

example is not surface true (due to the fact that sequences of CCiC nature do not occur because

high ranking faithfulness constraints preserve them but when CCC forces a vowel to be

inserted CiCC is preferred over CCiC A major contribution of OT has been focusing attention

on TETU of which many new cases have been found

Another advantage of OT over rule based generative approaches is its straightforward account of

what McCarthy (2001) calls homogeneity of targetheterogeneity of process A rule in rule

based approaches specifies the structure that it applies to (target) and the operation to be

performed on the structure (process) It has long been observed however that rules applying

different processes to the same target tend to occur cross-linguistically and within the same

language A rule based theory has no explanation as to why a structure should be a recurring

target In OT however the explanation is straightforward there is a markedness constraint

against the target but whether and how the target is repaired depends on interaction with other

constraints

McCarthy (2007) observes that the elements of a ranking argument are illustrated with a tableau

and that tableaux of two main types appear in the literature Each type has its usefulness for

certain purposes When the goal is to argue for ranking of constraints in a language then the

comparative tableau format of prince (2002) is used In this tableau each cell (row column)

indicates the number of violations if any of the constraint column incurred by candidate row as

shown in tableau (2) of the Yawelmani input Ɂilk-hin

input Ɂilk-hin Cu DEP Ɂilik-hin 1

Ɂilk-hin W1 LTableau (22) Comparative tableau for Yawelmani input Ɂilk-hin

Source McCarthy (20077)

In this kind of tableau every cell in a loser row has symbols W and L showing whether the

constraint favors the winner or the loser or no symbol if it favors neither For example in

tableau (2) the constraint Cu which bans syllabified consonants favors the winner because the

loser violates this constraint once while the loser violates it not at all DEP which ban

epenthesis of segments in an input favors the loser because the winner violates this constraint

and the loser does not The W and L annotations indicate the function of the constraint in the

system (McCarthy 2007)

The other type of tableau in the literature as observed by McCarthy (2007) is the violation

tableau of Prince and Smolensyky (2004) which is used when the goal is to show which

members of a given set of candidates are possible winners under different rankings of a given set

of constraints It allows for the observation of the difference in possible winners depending on

the ranking of the given set of constraints Tableau (3) shows how a violation tableau handles the

above Yawelmani input

input Ɂilk-hin Cu DEP Ɂilik-hin

Ɂilk-hin Tableau (23) Violation tableau for Yawelmani input Ɂilk-hin

In a violation tableau each violation of a constraint is indicated by an asterisk When a constraint

knocks a candidate out of competition the result is indicated by an exclamation mark Cells are

shaded when any violations that they may contain can have no effect on the outcome because

higher ranking constraint are decisive

This research employs the violation tableaux because the goal of the study is to show how

different candidate sets from EkeGusii and English are possible outputs under different rankings

of a given set of universal constraints Thus showing that languages differ in terms of how they

rank universal constraints and not in terms of rules some of which are language specific

OT like all other approaches to studies of this nature is not sort of weaknesses one challenges of

this approach is determining the range of candidates to be considered in an analysis given that

GEN has the potential of generating an infinite set of candidates (McCarty and Prince 1993

McCarthy 2007) This is because as McCarthy observes it is theoretically disastrous to overlook

a candidate that ties or beats the intended winner since the overlooked candidate has the potential

of undermining the entire analytical edifice

Equally challenging in this theory is to determine the most appropriate set of constraints and

their ranking to be used in the analysis of a given set of candidates given that constraints are

universal meaning that they affect all languages the difference being their different ranking in

different languages

CHAPTER THREE

RESEARCH METHODOLOGY

30 Introduction

This chapter describes the methods that were applied in carrying out this research It is organized

under the following sub sections research design research site study population sample size

and sampling techniques data collection procedures data analysis and ethical considerations

31 Research design

This study adopted both descriptive and explanatory research designs A descriptive design

attempts to show how the phenomenon under investigation is like Mugenda amp Mugenda (1999)

observe that descriptive research design determines and reports the way things are It attempts to

describe such things as possible behaviour attitudes values and characteristics In this study the

design allows generalized descriptions and characterization of the phonologicsl snd

morphological structures of of EkeGusii and English languages These kinds of descriptions and

characterizations allude to the analyses that are eventually carried out in subsequent objectives

as dictated by the theory in use (OT) This is in response to question one of the study

Explanatory research design on the other hand identifies the extent and nature of cause and effect

relationships It assesses the impact of specific changes on existing norms and processes Further

it analyzes situations to explain patterns of rlationships between variables (Zikmund Babin

Carr amp Griffin 2012) Explanatory design in this study allows for the assessment of the impacts

of EkeGusii phonological and morphological structures on the phonological and morphological

structures of English nouns borrowed into EkeGusii in response to questions two and three of the

32 Research site

The targeted research sites were Kisii and Nyamira counties which were purposively selected

because this is where most native speakers of EkeGusii (the target language) reside Through

simple random sampling Nyamira County was selected as the target research site It is in this

county that the accessible site (Nyagaachi Village) was selected for study This was carried out

as follows

Eighty (80) out of the one hundred and twelve (112) sub-locations within the county with the

desired characteristics (not within or near urban centers nor along ethnic boundaries) were

purposively selected with the assistance of the County commissionerrsquos office Nyamira County

This was in order to increase the possibility of selecting a sub-location with as many native

speakers of EkeGusii as possible with only first language (EkeGusii) competence As Weinreich

(1953) observes if a speaker is competent in more than one language heshe may attempt to

reproduce the borrowed morpheme with its original sounds while the monolingual speaker is

more likely to force the loan word to conform to the target language phonetic and phonemic

pattern From the selected sub-locations Enchoro sub-location was randomly selected Out of the

seventeen (17) villages of the sub-location (see appendix viii) Nyagaachi Village was randomly

selected thus becoming the research site of this study

33 Study populations

There were two types of population in this study population of the participants in the study and

the population of EkeGusii loan nouns from English The population of participants constituted

all the native speakers of EkeGusii in Kisii and Nyamira Counties as its target participants

population who according to the Kenya National Bureau of statistics (2009) are 2205669

However the accessible population of the study was the number of EkeGusii native speakers in

the selected study area in (32) above According to the Kenya Bureau of Statistics (2009) a

rural village in these counties has an average of 500 people distributed across ages This was the

target population of the study It is from this population that a sample was selected to provide

All the English nouns borrowed into EkeGusii constituted the second type of population

Available literature indicate that the population of English nouns borrowed into EkeGusii is not

known This study therefore treated all the English nouns borrowed into EkeGusii as its target

loan nouns population

Secondary data was also used in this study The main source of secondary data included library

study in which existing literature (books thesesdissertations dictionaries and journals) were

reviewed It is this type of data that was used to describe the phonological and morphological

structures of EkeGusii and English languages in response to question one of the study

34 Sample size and sampling techniques

This sub-section addresses the characteristics and the size of the sample of the interviewees and

English the nouns borrowed into EkeGusii and the sampling techniques that were employed to

get the samples The interviewees were adult native speakers of EkeGusii with knowledge of a

given semantic domain by virtue of their training or experience An adult in the Kenyan context

is an over eighteen (18) years old person But in this study the preferred age is over sixty

following Mecha (2013) observation that such a person has been widely exposed to the use of

language in various social contexts and therefore is competent enough to provide the required

data The sample size of interviewees was selected as follows An interviewee was purposively

selected from each of the semantic domains of borrowed nouns identified Thus thirteen(13)

interviewees were selected given that there were thirteen (13) of such domains This was the

sample size of the interviewees

A total of 349 English nouns borrowed into EkeGusii were collected from the interviewees (see

appendix (v) All these nouns constituted the sample size of the study This is because the nouns

could not be sampled any further because first their number was fairly small and secondly

sampling them could leave out some which could be used to explain certain phonological and

morphological processes while those which could not describe some processes could be

sampled Thus the nouns were selected purposively to describe and explain a process when and

where it occurs

35 Data collection procedures

In addition to native speaker intuition data in this study was collected from the interviewees

through semi- structured interviews (see appendix ii) These interviews were based on an

interview guide a list of questions based on the various domains of life (see appendix iii)

Interviewee responsesesnarratives were recorded by a voice recorder and later transcribed using

a raw data recordingtranscription form (see appendix iv)

36 Data analysis

Data analysis according to Mugenda and Mugenda (1991 203) is the process of bringing order

structure and meaning to the mass of information collected It seeks to make general statements

on how categories or themes of data are related Data in this study were in form of texts and were

of two types phonological and morphological Analyses of the data in this study were carried out

within the tenets and principles of the constraint based Optimality Theoretic (OT) framework

361 Procedure data analysis

English nouns borrowed into EkeGusii were analyzed against EkeGusii and English

phonological and morphological constraints rankings in order to account for the various

phonological and morphological changes observed phonemically since constraint ranking

between any two languages differ This according to the theory is carried as follows INPUTS

are subjected to the GEN component of the theory which generates an infinite set of candidates

The candidates are then subjected to the EVAL component which using the CON component

(ranked on a language specific basis) assesses and selects the most harmonic candidate

depending on the grammar in question The selected candidate becomes the OUTPUT of the

grammar This is illustrated by figure(2)

CANDIDATES CONSTRAINTS

INPUT GEN OUTPUT

Figure (2) process of OT realization of output

Phonological and morphological forms of English nouns borrowed from EkeGusii English

nouns and EkeGusii nouns served as inputs to yield outputs Constraints were ranked on

language input basis All these were aimed at establishing the constraint ranking that the

borrowed nouns adapted that of English or that of EkeGusii

Acoustically the vowels of the two languages were analyzed using the Praat computer software

This was in order to determine the acoustic nature and differences between the vowels of the

languages thus establishing the general direction of change

37 Ethical considerations

According to Kumar (1999) ldquoin every discipline it is considered unethical to collect information

without the knowledge of the participants in a research and their expressed willingness and

informed consentrdquo Seeking informed consent is probably the most common method in medical

and social research (Bailey 1978) It against this background the consent of the participants

(appendix ii) and government authorities were sought This was after describing to them the aims

and objectives of the study This was to ensure that the rights of the participants were guaranteed

Other ethical issues that were held include maintaining confidentiality of information by

participants avoiding bias in sampling especially of the participants use of appropriate

methodology correct reporting and appropriate use of information (Kumar 1999)

A Research Clearance Permit and a Research Authorization Letter were obtained from the

National Council for Science and Technology (appendices x and xi respectively) to ensure that

the study was conducted within proper ethical confines as required by law

CHAPTER FOUR

DATA PRESENTATION ANALYSIS RESULTS AND DISCUSSIONS

40 Introduction

This chapter is organized into three major sub-sections 41 presents a theory-neutral generalized

description of the phonological and morphological structures of EkeGusii and English while 42

and 43 presents phonological and morphological analyses respectively of the changes involved

in the nativization of English nouns borrowed into EkeGusii within the Optimality Theoretical

framework The chapter focuses on the phonological and morphological features of EkeGusii

and English lanuages that are affected in the process of nativization

41 Phonological and morphological structures of EkeGusii and English

In this sub-section descriptive generalizations of the phonological and morphological structures

of EkeGusii in comparison with those of English are given with the aim of providing the

structural differences and similarities which ultimately engender phonological and

morphological nativization in EkeGusii as analyzed in sub-sections 42 and 43 respectively The

descriptions allude to the tenets of Optimality Theory which provides that a descriptive

generalization is the essential intermediate step between data and analysis and that good

descriptive generalizations are accurate characterizations of the systematic patterns that can be

observed in the data Therefore according to the theory proceeding straight from the data to the

analysis without taking time to formulate an accurate descriptive generalization is never a good

idea The descriptive generalization mediates between the data and the analysis it is what the

analysis is an analysis of (McCarthy 200834) Data described in this sub-section is secondary as

gathered from existing literature including published books dictionaries theses and journals

411 EkeGusii and English vowel systems

Available literature indicate that few studies have been conducted in EkeGusii language

especially in the area of phonology Whitley (1960) which is among the pioneering studies in

the language lists seven basic vowels which are described by Cammenga (200239) as repeated

in chart (1) i e ɛ a ͻ o uHigh + + - - - + +Mid - + + - + + -Back - - - - + + +ATR + - - + - - +Chart (1) EkeGusii vowels Other studies carried out in the language (Osinde 1986 Ongarora 1996 2009 amp Bosire 1993

among others) have also confirmed that EkeGusii has seven relatively pure vowels as described

in chart (1) above This is further supported by a survey carried out by the University of

California in 1984 (Los Angles Phonological Segment Inventory Database) which found out that

most Bantu languages surveyed have between five and seven vowels (Odlin 2000) EkeGusii

falls within the seven vowel system The survey above places a seven vowel language system in

the cardinal vowel diagram designed by Jones (1956) as illustrated by chart (2)

Front Back High i u Mid high e o Mid low ε ͻ Low a

Chart (2) EkeGusii vowel trapezium According to Johnson (1997) the short vowels in chart (2) may occur both as short and long

depending on the environments they find themselves in as illustrated by (3) below (3) Short and long EkeGusii vowelsVowel Examples of words Gloss i siba sia tie

ii siiba siia sipe embori embori goatee embeera embeera graveεε orobeere ͻrͻεεrε titaa abaana aaana childrenͻ omoeto ͻmͻεtͻ trapͻͻ omoonia ͻmͻͻnia selleru ekeguuru ekeγuuru small potuu ebiguuru eiγuuru small pots(3) shows that every short vowel has its long counterpart making the total number of these

relatively pure EkeGusii vowels fourteen and not seven as has previously been described This is

indeed the position taken by this study

According to Johnson (1997) and Mihalicek and Wilson (2011) in speech the resonant

frequencies of the vocal tract or the frequencies that resonate the loudest are referred to as

formants It is these formants that are seen as peaks in a spectrum In their articulation vowels

produce several formants However as Mihalicek and Wilson point out the first three of the

formants labeled F1 F2 and F3 respectively are the most informative in speech The values of

these formants differ from vowel to vowel which leads to the distinction that is heard between

vowels and other sounds Spectrograms in figure (2) of the seven EkeGusii vowels were

produced by a male adult native speaker of EkeGusii

Time (s)0 0136

quency (

5000 ε

Time (s)0 007152

5000sound ee

Time (s)0 03045

5000EkeGusii sound a

Figure (3) Spectrograms of EkeGusii vowels [a e ε i ͻ o u] collected from a native speaker The vowels in figure (3) can be listed against their F1 and F2 as in table (2) as followsTable (2) F1 and F2 frequencies in hertz (Hz) of EkeGusii vowels Vowel F1 F2i 540 2450e 730 2250ε 830 2100a 900 1850ͻ 750 1250o 590 1150u 520 1100

Time (s)0 008388

quency (

5000sound ou

Time (s)0 007639

5000sounnd i

Time (s)0 1189

0594285714u

Time (s)0 007639

5000sounnd ii

Time (s)0 0136

ency (

5000ͻ

2500 2000 1500 1000 500

To come up with EkeGusii vowel trapezium as that in chart (2) above the vowels in table (2) are

plotted by frequencies of their first two formants as in figure (4) below

Figure (4) Plot of F1 against (F2) formant frequencies in hertz (Hz) of EkeGusii vowelsThis plot shows that the first formant corresponds inversely to the height dimension (high

vowels have low F1 and low vowels have a high F1) and the second formant corresponds to the

advancement (frontback) dimension (with front vowels having a high F2 and back vowels

having a low F2)

In comparison to EkeGusii English has a relatively large number of vowels which like those of

EkeGusii are either relatively pure or clearly gliding in nature (Cruttenden 2011 Roach 1983

OrsquoConnor 1967) among others identify the following vowels as adapted in (4)

4) English vowels

a) pure vowels

Vowel Examples of words

i heed feel bead

ɪ hid fill bid

e head fell bed

ᴂ had bad mad

ɑ hard bard par

ɒ hod bod

ͻ hoard fall board paw

ʊ hood full

u would fool booed pooh

ʌ but cut hut

ɜ heard fur bird pur

ǝ accept father

b) Dipthongs

eɪ fail bayed pay

aɪ hide file bide pie

ͻɪ foil boy

ǝʊ hoe foal

aʊ howersquod foul bowed pow

ɪǝ beard beer

eǝ haired bared pair

ʊǝ poor

c) Triphthongs

aɪǝ fire tyre choir society buyer

aʊǝ our shower flower coward nowadays

eɪǝ player greyer layer conveyor

ǝʊǝ mower slower

ͻɪǝ employer enjoyable buoyant joyous

(4) shows that English vowels like those of EkeGusii are grouped into categories based on their

quality (Cruttenden 2011 Gussenhove amp Jacobs 2011 Ladefoged 2001 Roach 1983 and

OrsquoConnor 1967)

Pure vowels remain constant and do not glide (that is move from one vowel to another) during

their production The vowels can either be long or short in nature as can be observed in data (4a)

above Long vowels marked by one vowel symbol and a length marker of two dots () are those

which take a relatively longer period to produce for example u Short vowels on the other

hand are marked by one vowel without any length marker they take a relatively shorter period to

produce (Gussenhoven amp Jacobs 2011 Cruttenden 2011 and Ladefoged 2001 Roach 1983)

The English pure vowels are listed in (5)

5) English pure vowels

Vowel Description

i long spread or non- rounded front high

ɪ short non-rounded front and high vowel

e short non-rounded front close-mid and open-mid

ᴂ short non-rounded front open-mid

ʌ short non-rounded central open

ɑ long non-rounded open central

ɒ short rounded back open

ͻ long rounded back mid

ʊ short rounded back mid-close

u long rounded back close

ɜ long non-rounded central mid

ǝ short non-rounded central neutral

The English vowels are many in number as compared to those of EkeGusii Thus there are a

number of vowels found in English but not in EkeGusii Both English and EkeGusii pure vowels

are characterized by length (6) compares the English pure vowels with those of EkeGusii

(6) Comparison between English pure vowels and EkeGusii vowelsEnglish pure vowels EkeGusii vowels

ᴂ ε

ͻ ͻͻ

(6) shows that while English has twelve pure vowels EkeGusii has fourteen Both EkeGusii and

English vowels have both long and short vowels EkeGusii length here is presented by doubling

of the affected vowels

The pure vowels in the two languages however are not the same especially in quality and

production Acoustically therefore even though these vowels share the same IPA symbols such

as [i] [e] [ͻ] and [u] they are different as illustrated by tables (3) and (4) of F1 and F2 of the

languages

Table (3) F1 and F2 frequencies in hertz (Hz) of EkeGusii vowels [i e ͻ u] repeated from

Table (2) above

Vowel F1 F2

i 540 2450e 730 2250ͻ 750 1250u 520 1100

Table (4) F1 and F2 frequencies in hertz (Hz) of the English vowels [i e ͻ u] taken from

spectrograms in figure (5)

Vowel F1 F2

i 280 2250e 400 1920ͻ 590 850u 310 890

Figure (5) Spectrograms of 8 British English vowels(Source Ladefoged and Keith 2001175)Tables (3) and (4) show that the frequencies of the first and second formants of the given vowels

are different For example while the formants of the English vowel i are 280 and 2250 for F1

and F2 respectively the formants of the same vowel in EkeGusii are 540 and 2450 respectively

Thus the acoustic nature of the vowels in these languages are significantly different and

therefore are heard differently

As pointed out above there are gliding vowels in English These according Ladefoged (2001)

Clark Yallop and Fletcher (2007) Gussenhoven and Jacobs (2011) and Cruttenden (2011)

among others are sequences of vocalic elements which form a glide within one syllable Those

made up of two such elements are called diphthongs as in (8b) while those made up of three are

called triphthongs as in 8c) above

Diphthongs have a first element (the starting point) and a second element (the point in the

direction of which the glide is made) According to Roach the RP diphthongs have as their first

element sounds in the general region of [ɪ e a ʊ] in which there are the diphthongs ɪǝ eǝ aɪ

aʊ ǝʊ and for their second element [ɪ ʊ ǝ] where there are the diphthongs ǝɪ eɪ ͻɪ The

following figure adapted from Roach (198319) gives a summary of the English diphthongs

DIPHTHONG

centring closing

ending in ǝ ending in ɪ ending in ʊ

ɪǝ eǝ ʊǝ eɪ aɪ ͻɪ ǝʊ aʊ

Figure (6) English diphthongs

But as pointed out above EkeGusii diphthongal combinations unlike those of English are

determined by vowel harmony and not the position of the first element as pointed out above

English unlike EkeGusii as pointed out above has triphthongs A triphthong is a glide from one

vowel to another and then to a third all produced rapidly and without interruption (Roach 1983)

Phonologists such as Roach (1983) and Cruttenden (2011) view a triphthong as being composed

of a closing diphthong with ǝ added on the end This means that a triphthongal vowel is

composed of three constituent vowels The five English triphthongs according to Roach (1983)

are composed of the five closing diphthongs- eɪ aɪ ͻɪ aʊ and ǝʊ- with an added ǝ Thus there

are five triphthongs in English as shown in (4c) above

Comparatively EkeGusii unlike English does not have triphthongs Vowel gliding in EkeGusii

ends at the second consonant

4111EkeGusii vowel harmony and disharmony

EkeGusii like some other languages is characterized by vowel harmony According to Sasa

(2009) vowel harmony is a phonological occurrence in which vowels in a certain unit (such as a

word) agree with a certain other vowel (such as a vowel in the first syllable of a word or a vowel

with a certain feature specification) Archangeli and Pulleyblank (2007) observe that a harmony

system demands that two or more segments which are not necessarily adjacent must be similar in

one way or another in terms of features The opposite of vowel harmony is vowel disharmony A

number of types of vowel harmony have been identified and discussed For example Rhodes

(2010) mentions the following tongue root harmony height harmony palatal harmony rounding

harmony and labial harmony EkeGusii language is characterized by vowel height harmony

(Ongarora 1996)

Vowel height harmony according to Oden (1996) is a characteristic of most Bantu languages

Phonology Oden observes that while any vowel in these languages can appear in the first root

syllable of a word affixes draw from a more restricted vowel inventory Typically affix vowels

distinguish only three vowels [a] and a frontback pair not of the third degree of height [i u] [ɪ

ʊ] or [e o] but not [ɛ ͻ] depending on the language The final vowel affix is usually drawn from

[i ~ ɪ] for negation [ɛ ~ ɪ] for subjunctive and [a] otherwise This is true of EkeGusii language as

discussed by Rhodes (2010) In discussing height harmony Rhodes (2010) notes that in

EkeGusii in addition to [a] high vowels block height harmony as demonstrated by (7)

7) EkeGusii vowel height harmony and disharmony a) tͻ-γɛɛnr-ɛ lsquolet us gorsquo Omo-te lsquotreersquo b) ͻ-rɛntir-e lsquoheshe has broughtrsquo e-ɳuͻm -ͻ lsquomarriagersquo ti-to-ko-ɳa-koβa-tɛβ-i lsquowe will not be telling themrsquo

EkeGusii has two mid vowel heights High mid and low mid (Ongarora 1996 Rhodes 2010

Anyona 2011 amp Cammenga 2012) This is illustrated by chart (2) above (7) shows that affix

mid vowels agree in height with root mid vowels For example in the word [tͻ-γɛɛnd-e] the root

vowel is the mid vowel [ɛ] while the prefix vowel is the mid vowel [ͻ] These two are in height

harmony However as (7b) shows if a non mid vowel intervenes between an affix mid vowel

and the nearest root mid vowel agreement is blocked For example the affix vowel in the word

[e-ɳuͻm-ͻ] is high mid rather than lower mid which would match the vowel in the root The

height of the first vowel can be attributed to the presence of [u] a high vowel between the two

mid vowels The vowels [e] and [ͻ] in this word are in vowel height disharmony (8) below gives

more examples of EkeGusii height vowel harmony and disharmony8) EkeGusii vowel height harmony and disharmonyi) Vowel Harmony[ͻmͻ-ɛt-ͻ] trap[ͻmͻ-ɳɛn-ɛ] owner[e-ŋgor-o] hole[omo-rem-i] farmer

ii) Disharmony|eke-suunt-e| [ eγe-suunt-e] darkness[Omo-ib-i] thief[ama-is-ͻ] eyes[omo-uk-ͻ] blind personAdapted from Bosire amp Machogu (2013)

In data (8i) the first syllables in the roots dictate that the prefix be in harmony (height) with it

For example in the word [e-ŋgor-o] the root vowel [o] is in harmony with the prefix vowel [e]

In (8ii) however this is not the case The first syllable vowels in the roots (which are either

[+HIGH] or [+LOW] dictate that the vowels be in disharmony with those of the prefix For

example in the word [eγe-suunt-e] the first vowel of the root syllable [u] dictates that the vowel

of the prefix be in disharmony with it instead of being in harmony that is [+MID HIGH] Three

types of harmony have been described in the literature total harmony opacity and transparency

(Sasa 2009) Sasa represents these schematically in (9)9) a) V1 V2 V3

[ F] [ F] [ F] (total harmony)

b) V1 V2 V3

[ F] [F] [ F](opacity)

c) V1 V2 V3

[ F] [ F] [ F](transparency)

(Where F represents any feature of the vowels such as [ATR] [LOW] and [ROUND] and the

Greek letters and the different values of the given features such as [+ ATR -ROUND

+HIGH])

In explaining the concept harmony the terms trigger and target are frequently used (Sasa 2009)

Sasa observes that the term lsquotriggerrsquo refers to the vowel with which all other vowels agree in

certain feature(s) while lsquotargetrsquo refers to the vowel(s) which agree(s) with the trigger in a given

harmony domain such as a syllable or a word In vowel harmony therefore it is targets which

harmonize with triggers

In total harmony represented schematically by (9a) all the vowels in a domain agree with the

trigger This is present in EkeGusii language as exemplified by the word [ͻ-mͻɳɛn-ɛ] lsquoownerrsquo

given in (9) above The trigger vowel [ɛ] in the root spreads the feature [+MID +HIGH] to both

the prefix and the suffix (the final vowel)

In opacity harmony (which contains an opaque vowel) the vowel adjacent to the trigger does not

agree with the trigger of the harmony In addition the final vowel agrees with the trigger of the

harmony This kind of harmony is equally present in EkeGusii as in the word |eke-suunt-e| gt

[eγe-suunt-e] lsquodarknessrsquo Here the opaque vowel [u] blocks harmony in the word except that

between the root and the final vowels The other two opaque vowels [a] and [i] trigger the same

behavior in EkeGusii

Transparency harmony contains a transparent or neutral vowel Here the medial vowel does not

agree with the trigger and the target it does not participate in harmony This is equally present in

EkeGusii language as in the word [a-ma-isͻ] lsquoeyesrsquo where the trigger vowel [ͻ] does not agree

with the medial vowel [i] or the target [a]

From the foregoing discussion on EkeGusii vowels it has been observed that EkeGusii has seven

vowels which can be classified as i e o u with advanced tongue root (ATR) and ɛͻ a with

retracted tongue root (RTR) As can be observed from chart (2) above only the mid vowels have

advanced and retracted counterparts as shown in (10)

10) Advanced and retracted tongue root mid vowel pairingATR RTRe ɛo ͻThe low retracted tongue root vowel a does not have a counterpart just like the high vowels i

u Like in other languages with seven vowels such as Yoruba (Pulleyblank 1996) only words

containing mid vowels show perfect harmony as illustrated by EkeGusii examples given (11)11) EkeGusii perfect harmony ATR RTResese esese lsquodogrsquo ɛsɛsɛ esese lsquocoughrsquoomoonto omoonto lsquopersonrsquo ͻmͻɛtɛ lsquotraprsquoookombe obokombe lsquohoersquo ͻmͻxɛrɛ lsquoLuorsquoAdapted from Bosire amp Machogu (2013)English is not characterized by this phonological process According to Shapiro (2015) English

lacks vowel harmony as a regular phenomenon

412 EkeGusii consonant system in comparison with English

In this section the consonant system of EkeGusii is compared and contrasted with that of

English Studies on EkeGusii phonology have identified a number of consonants For example

Cammenga (200253) has identified the consonants in chart (3)

Bilibial Alveolar (Alveolo-) Palatal Velar

Continuant β s γ

(Flapped liquid) r 57

Obstruent [b] t k

Affricate c

Nasal m n ntilde ɳ

Glide w y

[Cw] [Cy]

Chart (3) EkeGusii consonantsThis study makes the following observations about consonant inventory in chart (3) Firstly it

should be noted that Cammengarsquos (2002) inventory of EkeGusii consonants is an improvement

of Whiteleyrsquos (1960) inventory In Whiteleyrsquos (1960) inventory are the following consonants

which Cammenga does not include in chart (3) above [p] [ny] and [y(j)] As observed by both

Cammenga and Whiteley the voiceless bilabial stop [p] is only found in EkeGusii words

borrowed from languages in which the sound is present such as Kiswahili and English It can

therefore be concluded that the sound is not found in EkeGusii language except in ldquoone or two

idiophonesrdquo as suggested by Whiteley The idiophone suggested by Whitely would be the

emphatic form pi which means lsquocompletelyrsquo as illustrated by (12)

12) EkeGusii ideophone with the voiceless stop pi) ita pi ita pi lsquokill completelyrsquoii) geenda pi γɛɛnda pi lsquogo completelyrsquoiii) koora pi koora pi lsquofinish completelyrsquoAdapted from Bosire and Machogu (2013)This data shows that pi in the words emphasizes the given actionsSecondly Cammenga (2002) replaces ny with ntilde and names j a glide instead of a semi vowel

This study will use the IPA symbol ɳ to represent the palato-alveolar nasal instead of ny and

rename j an approximant instead of a glide

Thirdly following observations that Bantu languages do not have consonant glide sequences but

instead that the glides (approximants in this study) are realized as secondary articulations58

(Hayman amp Katamba 1999) what Cammenga includes as consonant glide sequences ([Cw] and

[Cy]) will not be included in the inventory in this study The approximant w will be excluded

altogether from EkeGusii consonant inventory meaning that it will only be treated as a derived

secondary consonant represented as ([C[w]])

Fourthly the pre-nasal stops [b] [d] and [g] the voiced alveolar fricative [z] and the voiced

palate-alveolar fricative [dʒ] like the secondary approximants described above will be treated as

derived consonants through homorganization and defricativization They are therefore not part of

the phonological system of the language This then means that they are equally treated as

secondary derivativations

Fifthly the affricates that Cammenga (2002) represents with the symbols c and dŽ are in this

study represented as the IPA symbols tint and dʒ respectively

EkeGusii consonants can now be represented as in chart (4)

Bilabial Alveolar (Alveo-) Palatal Velar

Continuant ɸ s γ

(tril) r

Obstruent [b] t k

[d] [g]

Affricate [dʒ]

Nasal m n ɳ ŋ

Approximant [w] j

Chart (4) EkeGusii consonant inventory Chart (4) shows that two new consonants have been added into the consonant inventory of

EkeGusii These areɸ voiceless bilabial continuant as in obuba oɸuɸa lsquofoodrsquo amaraba

amaraɸa lsquosoilrsquo abasaacha aɸasaatinta lsquomenrsquo and x voiceless velar continuant as in omogesi

omoxesi harvester agaanto axaanto lsquoa thingrsquo ensagara enzaɸara lsquolizardrsquo

Therefore this study concludes that EkeGusii has fourteen distinctive consonants in its

phonological inventory ɸ s γ x r t k t m n ɳ ŋ and j and six

phonetic derivatives [z] [b] [d] [g] [dʒ] and [w]

Phonologically voiced EkeGusii consonants seem to occur with the mid-high vowels e and o

(with the feature [+ATR]) while the voiceless ones occur with the mid-low vowels ε and ͻ

(with the feature [+RTR]) The rest of the vowels occur without such restrictions This is vowel-

consonant harmony controlled y the feature [VOICE] (13) illustrates this observation

13) Occurrence of vowels with consonants in EkeGusii e and o (ATR) ͻ and ɛ (RTR)ebando eβando lsquomaizersquo oboba ͻͻa lsquomashroomrsquoegesanda eγesanda lsquocalabashrsquo etoigo εtͻixͻ lsquofloodsrsquoemondo emondo lsquogizzardrsquo omoeto ͻmͻεtͻ lsquotraprsquoAdapted from Bosire and Machogu (2013)

Whether a vowel occurs with a voiced or a voiceless consonant in EkeGusii seems to be

determined by whether the vowel is advanced or retracted tongue root This in fact is what is

responsible for the consonant and vowel harmonies that are observed in data (13) Words having

vowels with ATR demand [+VOICE] consonants while those with RTR demand [-VOICE]

consonants This is further exemplified by data set (11) above

As compared to EkeGusii there are 24 distinctive phonological units which are consonantal both

in terms of their position in the syllable that is phonologically and also in the majority of cases

in terms of how they are produced in vocal tract that is phonetically (Cruttenden 2011) These

consonantal phonemes are classified into two broad categories Obstruents (those articulations in

which there is a total closure or a stricture causing friction) This group is associated with a noise

component which accompanies their production They are further characterized by a distinctive

opposition between voiceless and voiced types The other category of consonants is that of

sonorants These are those consonants characterized by articulations in which there is only a

partial closure or an unimpeded oral nasal escape of air Such articulations are normally voiced

and frequently frictionless that is they are without the noise component of the obstruents This

class shares many phonetic characteristics with vowels

According to Chomsky and Halle (1968) obstruents and sonorants are features that classify

segments according to their noise component Those in whose production the constriction

impeding airflow through the vocal tract is sufficient to cause noise are known as obstruents

while those in which there is no noise component are known as sonorants The following English

consonant classes belong to the obstruent class bilabial plosives b p alveolar plosives t d

velar plosives k g palatal alveolar fricatives tint dʒ labiodental fricatives f v dental

fricatives θ eth alveolar fricatives s z alveolar fricatives int ʒ and glottal fricative h

Sonorants on the other hand are those sounds in which there is no noisy component in their

production This group has the following classes of consonants nasals approximants and

vowels Vowels having been described in the previous section the rest of these sonorants are

described as follows as nasals bilabial nasal m alveolar nasal n and the velar nasal ŋ

approximants the lateral approximant l post alveolar approximant r unrounded palatal

aapproximant j and the labial velar approximant w

The description of the English vowels and consonants in this research has relied on Cruntenden

(2011 pp 91-232) Details and clarification therefore can be verified The English consonants

described so far are presented in a manner and place of articulation chart as in chart (5)

Bilabial Labiodentals Dentals Alveolar Palate-alveolar

Palatal Velar Glottal

Plosive p b t d k gFricative f v θ eth s z ʒ hAffricative t dʒNasal m n ŊLateral lApproximant w r j

Chart (5) The English consonantsAdapted from Jones (1972 xvii)A number of observations about the consonant inventories of English and EkeGusii described in

this section can be made Firstly EkeGusii consonants are fewer as compared to those of

English while EkeGusii consonants are fourteen the English consonants are twenty-four (14)

below gives inventories of the consonants in English and EkeGusii(14) Inventories of English and EkeGusii consonantsEnglish consonants EkeGusii consonants p

b -f -v -θ -eth -t t d -m mn n- ɳ ŋ ŋk kg -- γ - xj jw - r rl -

dʒ -tint tints sz - - ʒ -h -The second observation is that while some of the consonants in the inventories are similar or are

the same featurewise others are not This is clearly captured in (14) which shows that EkeGusii

consonants ɸ ɳ γ x are absent in English while the English consonants p b f v θ eth d

g w dʒ int ʒ h are absent in EkeGusii

The third observation is that some consonants are shared at least in terms of phonetic features by

both phonologies These consonants are [t m n ŋ k j r tint s]

4121 EkeGusii consonantal processes

Like other languages EkeGusii phonology is characterized by consonantal processes The

processes described inthis section are those which affect EkeGusii phonology and therefore the

English loans in the language they may no affect English phonologyThese sub-section describes

these processes

41211 Voice dissimilation (Dahlrsquos Law)

This process according to Uffmann (2013) is found in a number of Bantu languages EkeGusii

is one of the languages characterized by the process Uffman defines Dahlrsquos law as a voicing

dissimilation process in which a prefix stop which is underlyingly voiceless is voiced if the stem

or subsequent prefix starts with a voiceless segment Guthrie (1967) observes that languages

which show the effect of this law are found within his zones E20 - E50 ndash F20 and G60 EkeGusii

language zoned E42 is within this range

It has also been argued that languages vary a great deal as to which consonants undergo the rule

which consonants trigger the rule and how the rule affects multiple targets within the same word

(Bennett 1967 and Davy and Nurse 1982) This means that different languages have different63

consonants which undergo the process different consonants acting as triggers in different

languages and different effects on targets in different languages (Bickmore 1998) Bickmore

observes that in EkeGusii there is evidence that Dahlrsquos law affects the dorsal stop [k] as (15)

below demonstrates

(15) The effect of Dahlrsquos Law on [k] in EkeGusii |ͻkͻ- kɛsa| ͻkͻ-لاεsa lsquoharvestrsquo |oko-koro| oko-لاoro lsquolegrsquo Source Bickmore (1998)

This data shows that the voiceless velar obstruent k in the prefixes ͻkͻ- and oko-

respectively are substituted for by the voiced velar obstruent γ in the roots ndashγɛs and γor

respectively Thus the k sound in the initial syllables does not assimilate the sounds in the

adjacent syllables as expected in most languages including English Instead it dissimilates as

shown This process is still quite productive in the synchronic phonology of EkeGusii

(Bickmore 1998) as exemplified by the class 15 prefix ko- in (16)16) Dahlrsquos Law in EkeGusii Word underlying form surface form gloss a) okoroota |ͻ-kͻ-rͻͻt-a| [ͻ-kͻ-rͻͻt-a] lsquodreamrsquo

okogoro |o-ko-kor-o| [o-ko-γor-o] lsquofootrsquookonywa |o-ko-ηw-a| [o-ko-ηw-a] lsquodrink

Adapted from Bickmore (1998) The dissimilation process in (16a) is from the voiceless obstruent stop k to a voiced obstruent

fricatives such as γ and the other way round in (16b)

41212 Prenasalisation and nasal homorganisation

Prenasalisation is the process which is responsible for the derivation of prenasalised consonants

This process according to Cammenga (2002) causes the nasal part of the prenasalised consonant

to become homorganic with the following consonantal element Thus both the nasal and the

consonantal elements involved in the process share place features of the consonantal element In

other words hormorganization is the process where the nasal element of the pre-nasalized

consonant becomes homorganic (they both share the place features with that consonant) (17)

adapted from Cammenga (2002 87) demonstrates this observation 17) EkeGusii nasal homorganisationInput ɛ-n- + -γͻkͻ lsquohenrsquoAffixation ɛnγͻkͻ Prenasalisation ɛnγͻkͻNasal homorganisation ɛŋγͻkͻ

(17) shows that the nasal n which is [+alveolar] becomes [ŋ] a [+velar] consonant a place

feature of the consonant γ This is nasal homorganisation This process affects all nasal

elements of all prenasalised consonants in EkeGusii (Cammenga 2002) Thus |m+| rarr [mb]

while |n+r| rarr [nd] Thus it can be argued that the nasal plus consonant as given here produces a

secondary consonant such as [mb] which as will be observed in section (4113) below is

secondary realization which is treated as a single consonant and not a cluster

41213 Declusterization of nasal consonant (NC) and consonant glide (CG) clusters

This study argues that there are no consonant clusters of any nature in EkeGusii Thus

underlying nasal consonant and consonant glide clusters are declusterized in EkeGusii surface

forms This is in agreement with Hyman and Katamba (1999) who observe that Bantu languages

do not have consonant clusters To Hyman and Katamba the only combinations that seem to be

clusters of consonants are those of the nasal consonant (NC) consonant glide (CG) and nasal

consonant glide (NCG) This is the position taken in this study that EkeGusii does not have

obvious consonant clusters What seems to be nasal consonant and nasal glide clusters are in fact

secondary articulations motivated by the hormorganization process discussed in the previous

sub-section These nasal consonant and nasal glide secondary realizations are what the study

refers to as declusterization

Hyman and Katamba (1999) identify two kinds of consonant clusters that are of significance in

the phonology of Bantu homorganic nasal consonant sequences also called pre-nasalized

consonants discussed above and consonant glide sequences (CG) These two at times overlap to

produce a nasal consonant glide (NCG) cluster as illustrated by (18)

18) EkeGusii nasal consonant glide (NCG)

Underlying form surface form gloss

|n-βu-ate| rarr [mbwate] lsquohold mersquo

[m b w a t e]

Adapted from Katamba (1999)

(18) shows that the underlying nasal |n| is homorganized to [m] which in turn assimilates the

consonant |β| a continuant to [b] a stop Further the underlying vowel |u| which is high

rounded is assimilated to the glide [w] an equally rounded approximant by the vowel [a] which

is low This is for ease of articulation (Katamba 1993) (19) gives further examples of consonant

glide sequences yielding hormorganization

19) EkeGusii consonant glide hormoganization

i) Input buata lsquoholdrsquo

Output [wata]

Hormorganization process

βuata rarr [wata]rarr [βwu-ata]

ii) Input sieka lsquoclosersquo

Output sjeka

sieka rarr [sjeka]rarr [sjeka]

where Cw and Cj are secondary articulations

Adapted from Hyman and Katamba (1999)

There is enough evidence in support of the consonant glide hormoganization argument advanced

here as elsewhere In LuGanda for example when two vowels are adjacent the first vowel is

deleted unless it is high (in which case it becomes a glide [w] or [j] (Katamba1993) Similarly

in Emai if two vowels are contained in lexical morphemes following one another and that the

vowel in the first morpheme is high [i] or [u] the high vowel changes into homorganic glide of

the appropriate place of articulation (McCarthy 2007) as shown in (20) below repeated from

McCarthy (20079)

20) Emai consonant glide hormorganization

(i) ku ame rarr [Kwame] lsquothrough waterrsquo

(ii) fi ͻpia rarr [fjͻpia] lsquothrow cutlassrsquo

In (20i) the high vowel u hormoganically changes to the labial consonant glide w while in

(20ii) i changes to the palatal consonant glide j EkeGusii consonant glide hormoganization in

(19) above behaves the same way as the Emai hormoganization in (20) In lsquosiekarsquo in (19) for

example the high vowel i which is adjacent to the vowel e changes to the homorganic

consonant glide [j]

The secondary articulations as in Cw and Cj in (19) above advance the argument that instead of

treating a sequence as a consonant glide (CG) it is occasionally treated as a secondary

articulation on a single consonant [wata] and [sjeka] respectively This means that the consonant

here is one (the primary one- underlined which is accompanied with a secondary one which is a

semi vowel- superscripted) Similar arguments have been advanced by Hargus and da Conceicao

(1999) who propose that Ronga language (spoken in Mozamique) has distinctively labialized

consonants for example the nasal consonant [n] in the word [nwala] lsquofingernailrsquo rejecting a

cluster analysis on the grounds that there are no any other onset clusters in the language

Similarly Otterloo (2011) treats potential clusters of the type [Pj Kw] in Pahari language

(spoken in in Northeastern Parkstan) as violating secondary articulated palatalized and labialized

consonants [Pj and Kw] respectively

Following the foregoing observations this study argues in support of the view that EkeGusii

language does not have consonant clusters Instead it has secondary realizations in cases of

consonant glides as (21) further illustrates

21) Ekegusii consonant glides homorganization as secondary articulations

Word underlying form surface form Gloss

a) rwana ru-ana [rwana] fight

b) kwani ku-ani [kwani] greet

c) chwei tintu-eri [chweri] saw

d) etia e-ti-a [etja] pass it

e) berja βeri-a [βerja] boil

f) tjana ti-ana tjana swear

Example (21a) can be represented as in figure (7)

rwana instead of rwana

c vcv c cvcv

Figure (7) Articulation of complex glides in EkeGusii

Here the realization rwana is treated as ungrammatical because as has been argued before it

allows a cluster of consonants which is against Bantu phonology which disallows consonant

clusters

In syllabic nodes the syllables in (figure 7) above will be represented as in figure (8)

a) rwana σ σ σ

C V V C V

rw u a n a

b) rwana σ σ

cc v c v

rw a n a

Figure (8) EkeGusii consonant glide syllabic nodes

Figure (8a) shows that the realization of the given word is grammatical in that it does not have

any consonant cluster while realization (8b) is ungrammatical because it contains a consonant

cluster which is disallowed in Bantu phonology

Herbey (1986) and Downing (2005) pose two questions about nasal consonant sequences in

Bantu phonology They wonder if the sequence is a single segment or a cluster and if it is a

cluster how the given components are syllabified These are the same questions that this research

sought to answer

The reason why NC sequences such as [nt] and [nd] are treated as two segments which is rare is

that they are bi-morphemic arising by joining of an autonomous nasal (a consonant) with

another consonant For example in Matumbi language (spoken in Tanzania) the sequence [mb]

as in the word [mbajite] ldquoI saidrdquo derives from ldquonitbajiterdquo which is optionally realized as

[nimbajite] for ease of pronunciation (Herbey 1986 and Downing 2005) However the reason

why these NC clusters may not be treated as two segments especially in Bantu languages (which

favours the arguments advanced in this study) is that this would favor languages (such as

EkeGusii) with a typology of uncommon syllable structure such as onset and coda clusters

which violate the sonority sequencing principle (Sievors 1981 amp Clements 1990)

To avoid treating and calling NC sequences consonant clusters linguists employ the term

ldquoprenasalised stopsrdquo (Hearth 2003) According to Hearth Makaa a Bantu language spoken in

Tanzania for example has twenty-two simple consonants and eight prenasalized stops Equally

Alnet (2009) lists a series of pre-nasalized consonants in Shimaore language

Following the foregoing observations and arguments this study argues that EkeGusii has pre-

nasalized stops and other consonants and therefore no NC clusters in its syllable structure (22)

below gives the four pre-nasalized consonant stops in EkeGusii

22) EkeGusii prenasalised consonant stopsPrenasalised consonant Example of word Gloss

a) |n+b|gt mb engombe [ɛŋͻmbɛ] lsquocowrsquo

b) |n+r| gt nd enda [enda] lsquostomachrsquo

c) |n+t| gtnt egento [eγento] lsquothingrsquo

d) |n+k|gtŋg egechanga [eγetintaŋga]lsquowirersquo

In (22) the NC lsquoclustersrsquo (underlined) are treated as one consonant In other words there are no

consonant clusters in essence For example (22b) can be represented syllabically as in figure (9)

enda e nda

V NC V

e nd a

Figure (9) Nasal consonant syllabification in EkeGusii

The first syllable is made up of only the syllable nucleus which is allowable in this language as

in many other Bantu languages The second syllable it is argued is made up of a pre-nasal

consonant a consonant proper (and not two consonants) and a vowel Thus it has an onset a

prenasalised consonant and not an NC cluster

Clements (1978) observes that there is vowel lengthening before NC clusters in most Bantu

languages as illustrated by (23) adapted from Katamba (1989)

23) Ekegusii NC clusters

Word pronunciation gloss

a) omoonto omoonto person

b) ebaando eβaando maize

c) engombe εŋgͻͻmbɛ cow

d) eyaanga ejaaŋga dress

This data shows that the vowel before every nasal consonant is doubled (lengthened) For

example in omoonto in (23a) the vowel o in the prefix omo- is doubled so that it becomes

the nucleus of the initial syllable of the root nto Clements (1978) observes that such

lengthening regularly holds in many Bantu languages including Yao Hehe Sukuma and Kuria

spoken in Tanzania and Kikuyu Luhyia Kuria spoken in Kenya and many others

The assumption according to Clements (1978) is that a pre-consonantal nasal has a special

prosodic status that is dominated by a vowel rather than a consonant This normally results in

syllabification of the nasal into the coda of the preceding syllable but the fact that syllables

should not be closed (Prince and Smolensky1993) is taken to argue against positing nasals in the

coda position or having closed syllables The syllable is therefore syllabified in the onset of the

following syllable which leads to compensatory lengthening of the preceding vowel by re-

association of the standard timing unit as illustrated by (24)

24) EkeGusii compensatory lengthening of vowels

Input Omoonto lsquopersonrsquo

i) moonto nasal as Coda

ii) moonto nasal as Onset

iii)moonto nasal as syllabic consonant

This data shows that it is (24ii) which necessitates compensatory lengthening This argument

depends on the assumption that the nasal in the vowel NC sequence must be in non-linear

analysis (Clements 1986) Here the pre-nasal consonant lengthening is treated as compensatory

lengthening coming from the fact that the nasal is deprived of its vowel slot because it is moved

into the onset slot in the word and so a vowel must come in to fill the empty vowel space left by

the nasal This is demonstrated by Figure (10)

CVV CV C CVVC

[monto] [mo-nto] [moonto

Figure (10) EkeGusii vowel lengthening

Figure (10) indicates that the vowel [o] moves in (in 3) to fill in the gap left behind (in 2) by the

nasal [n] which is in the onset position (in 2) having moved from the coda position (a mora) (in

1) leaving behind an empty slot necessitating vowel lengthening This is presented on a syllable

node as figure (11)

σ σ σ σ

micro micro micro micro micro micro

m o n t o m o n t o

Figure (11) EkeGusii compensatory lengthening

The phonological evidence in support of the fact that the nasal in NC combination is Onset is

that in most languages most syllables are open that is syllables normally end in vowels (Kager

1999) However the phonotactics of English as will be discussed in sub-section (4113) do

allow consonant clusters It also allows closed syllables It can therefore be argued that while

EkeGusii does not recognize NC and NCG sequences as consonant clusters English does

41214 Defricativisation

Defricativization is another process that is caused by Prenasalisation Here according to

Cammenga (2002) if the consonantal element in the combination that is prenasalised is a voiced

continuant it loses the [+CONTINUANT] feature In other words it becomes [-

CONTINUANT] This Cammenga observes is accounted for as rightward spreading of [-

CONTINUANT] specification of the nasal to the consonantal element This process is described

thus Voiced continuants are turned into voiced obstruents whenever they are prenasalised In

EkeGusii β is turned into[b] γ into [g] and r into [d] as illustrated by (25)

25) EkeGusii defricativisationi) β rarr[b]

input e-n- + βori lsquogoatrsquoaffixation | e-n-βori|Prenasalisation enβoriNasal homorganisation [embori]

ii) γ rarr [g]input ɛ-n- + γͻri lsquoropersquoaffixation ɛnγͻriPrenasalisation ɛnγͻriNasal homorganisation ɛŋgͻri

iii) r rarr [d]input e-n- + raγeraaffixation enraγeraPrenasalisation enraγeraNasal homorganisation endaγera

Adapted from Cammenga (2002)

This data shows that whenever a voiced continuant obstruent is adjacent to a nasal it loses its

[+CONTINUANT] feature and becomes [-CONTINUANT] in other words it is defricativised

This confirms the fact that EkeGusii does not have the stops that are the end products of

defricativization ([b] [g] and [d] respectively

Comparatively defricativization is not a characteristic of English phonology as in EkeGusii This

can be explained by the fact that nasals plus consonants in English can be treated as consonant

clusters since the language allows consonant clusters as discussed in (4113) below Like in

EkeGusii however nasal consonant homorganization characterizes English as illustrated by (31)26) English nasal consonant homorganizationi) ink rarr iŋkii) tomb rarr tumIn (26i) the consonant k homorganizes n to [ŋ] while in (31ii) b disappears as a result of

being hormoganized to [m]

41215 EkeGusii Nasal re-syllabification

Ferguson (1963) Hyman (1985) and Nasukawa (2004) observe that syllabic nasals which are

found in languages such as Pali Japanese and many Bantu and Ogoni languages exhibit both

consonantal and vocalic characteristics in terms of their tonal properties and syllabic distribution

This is true of EkeGusii language According to Cammenga (2002) whenever prenasals occur

word initially their nasal elements may optionally become syllabic and bear tone This tone may

or may not be distinct from that of the next tone bearer that is the next syllable Such changes

may occur in word initial position only Nor does it seem to be limited to prenasals only In

explaining the nasal resyllabification process in this section this research will in the process

explain other rules which according to Cammenga and indeed this study are presupposed by the

process In fact Cammenga simply refers to the various processes which finally lead to

syllabification as delinking rule

Though viewed as optional occurrence in word initial positions nasal syllabification is a

common process especially in Bantu languages In cases where a nasal is followed by a

consonant syllabification takes place as exemplified by (27) adapted from Cammenga

(200290)

27) a) nasal syllabification in word initial position

n-to- taatint -ɛ

F-1p-fetch-FV

[ntotaatinte]

lsquoWe will fetch (water) todayrsquo

b) in- mo- taa ts- e

Fndash 2p1-fetch-fv

[mmotaatinte]

lsquoYou will fetch (water) todayrsquo

c) in-a-taatint-e

F-3p-fetch-fv

[mbataatinte]

lsquoThey will fetch water todayrsquo

Syllable nasals are underlined in (27) The data indicates that the nasal element is in the word

initial position There are cases where nasals may also be syllabified before vowels as in (28)

(28) nasal syllabification before vowels

Input -e- lsquoforgetrsquo

Suffixation -ee

Prefixation n-e-

Nasal resyllabification ne-

Pre-nasal i insertion ine-

Nasal velarization -iŋe-

Output [iŋee] ltingebegt lsquoforget mersquoAdapted from Cammenga (2002 90)

Data sets (27) and (28) are accounted for by word initial delinking rule which is exemplified by

figure (12)

[word[+nasal]] [+consonant]

Figure (12) EkeGusii delinking rule

Source Cammenga (2002)

This figure shows that the delinking rule optionally delinks in word initial position a pre-

nasalized consonant from the syllable ([]) to which it is attached This is what necessitates re-

syllabification This is because the delinked word initial nasal floats which by convention may

not be relinked to the following consonant figure (13) further exemplifies

[word [+Nasal]

Figure (13) EkeGusii nasal re-syllabification rule

This rule optionally assigns a nucleus to any floating word initial nasal This is nasal re-

syllabification The process of nasal syllabification starts with prenalization where word initial

nasals are pre-nasalized Pre-nasalization then triggers nasal homoganization in which a nasal

shares place feature with the consonant it precedes Then defricativization takes place where and

when applicable especially when the following consonant is a fricative(continuant) This is then

followed by the nasal de-linking process as in figure (12) above and finally re-syllabification as

in figure (13) This process is summarized in (29) as follows

29) EkeGusii nasal syllabification process

Input -γor- lsquobuyrsquo

Suffixation -γore

Prefixation n-γore

Prenasalisation nγore

Nasal homogenization ŋγore

Defricatirization ŋgore

Nasal delinking ŋ-gore

Nasal resyllabification ŋgore

Output [ŋgore]

This data show that the nasal consonant in the syllabified form forms the initial syllable of the

word in which it is initial This is after delinking itself from the syllable in which it is attached

This means that it does not form a consonant cluster with the consonant with which it occurs

This is illustrated by Figures (14) and (15) for the output in (29) above

N C V C V

ŋ g o r e

Figure (14) Nasal delinking leading to nasal syllabificationAdapted from Cammenga (2002)

CC V C V

ng o r e

Figure (15) Nasal consonant cluster forming part of a syllable

In figure (14) the nasal forms a syllable on its own it has delinked itself from the syllable to

which it is attached while in figure (15) it is part of the syllable it is attached thus forming a

consonant cluster which is not allowed in EkeGusii

Following the foregoing discussion and conclusions on nasal resyllabification which has mainly

drawn from Cammenga (2002) this study supports the argument that all nasals in EkeGusii are

syllabified This observation further supports the arguments that EkeGusii does not entertain

consonant clusters This is because the delinking of the nasal from the consonant with which it

occurs makes the nasal stand on its own as a syllable In EkeGusii like in most languages

vowels unlike consonants form syllables on their own Therefore the nasals in this study are

treated more as vowels as compared to consonants because they occupy vocalic positions in

syllables

Syllabification of nasals by delinking as described in this section does not characterize English

phonology English entertains consonant clusters and as has already been observed the nasal

plus consonant combination form a cluster Delinking a nasal from a consonant in English

creates non-syllabic structures as illustrated by (30)

30) Nasal consonant delinking in English

i) ink rarr iŋk

ii) ink rarr [iŋk] lsquoinkrsquo

(35ii) is ill formed because the nasal [ŋ] has been delinked from the consonant [k] creating two

unacceptable syllables (in English) instead of one as in (35i)

413 Comparison of EkeGusii and English syllable structures

The description of the syllable in this study is based on the typology of syllable inventories

originally stated in Jacobson (1962) and elaborated in Clements and Keyser (1983) and Prince

and Smolensky (2004) This is a typology based on syllable inventories attested across

languages It belongs to a class of substantive universals and includes the implicational relations

that hold among specific syllable shapes De Lacy (2006) gives a typology of different languages

syllable shapes repeated in table (5)Table (5) Typology of syllable shapes

Onset coda onset coda cluster Inventory Languages

cluster

O (C)CV(C)(C) Totonak

X (C)CV(C) Dakota

XO CV(C)(C) Klamath

X CV(C) TemierR X

O _ (C)VC Arabela

X CV Senufo

O O OO (C)(C)V(C)(C) EnglishX (C)(C)V(C) Spanish

X O (C)V(C)(C) Finish

X (C)V(C) Turkish

O XO - (C)(C)(V) PirahaX - (C)V Fijian

Adapted from De Lacy (2006 165)Key

R= Required O= Optional X= BannedX therefore means that

Codas are never requiredOnset clusters are never required

Coda clusters are never requiredOnsets are never banned

Generally Ekegusii has a (V) CV syllable structure (Cammenga 2002) Thus the language

is characterized with an open syllable structure and sometimes a single vowel word initially

as illustrated by (31)

31) EkeGusii syllable structure

Syllable Underlying Surface EkeGusii Gloss

CV CV |ketii| [γetii] getii field

a) CVCV βana| [βana] bana predictfore-tell

b) CVCVCV tɛrɛɛra [tɛrɛɛra] tereera sing for

c) VCVCV omote [omote] omote tree

d) VCVCV CV omoγori [omoγori] omogori buyer

e) VCVCVCV aaani [aaani] ababani prophets

Adapted from Bosire and Machogu (2013)

(31a) for example can be presented on syllable nodes as in figure (16)

Input ketii output [γetii] lsquofieldrsquo

C V C V

γ e t ii

Figure (16) EkeGusii syllable nodes for γetii

These syllable structures generally presuppose that syllables should have onsets and that

the consonants in the input are the same as the consonants in the output respectively

(Smolensky amp McCarthy 1993)

There are cases of single vowel syllables in EkeGusii This however is a case of onset

violation where a vowel begins in a word especially in nouns number and class marking

pre-prefixes and prefixes and in some single vowel words as illustrated by (32)

32) Single vowel syllables in EkeGusii

a) i) o- mo -te lsquotreersquo ii) e- mete lsquotreesrsquo Aug prefix root Aug prefix root 3sg 33 -tree (sing) 4pl 34 - tree (pl)

b) aaa lsquopluckrsquo (vegetables etc)The prefix omo- in (32ai) above marks the class of the noun lsquotreersquo that is class 3 and

number that is singular while the prefix eme- in (32aii) marks class four and plurality

Examples (32ai and (32b) above will be represented on a syllable node as in figures (17)

i) omote

vc vcv σ σ σ

v c v c v

o m o t e ii) a aa

vv σ σ

a aa Figure (17) EkeGusii syllable nodes for [omote] and [aaa]

EkeGusii syllable structure is comparable to that of English While EkeGusii is a (V)CV

language as shown in (31) the syllable typologies given in table (3) above show that

English is a (C) (C) V (C) (C) syllable type of language Thus Onsets Codas Onset

clusters and Coda clusters are all optional in English They may or may not occur

depending on the nature of the word as shown in (33)

33) Syllable types in EnglishWord syllable typecat kᴂt cvc boy bͻɪ cvstructure strᴧktintə cccvccv owe əʊ vAccording to Roach (1983) if the first syllable of a word begins with a vowel (and in

English any vowel may occur though ʊ is rare) the syllable is said to have a zero onset If

it begins with one consonant that consonant may be any except ŋ and ʒ which are rare

in this position

There are two types of two-consonant clusters in English that which begins with s as in

string sting sway and smoke In this case the s is pre-initial while the other consonants

eg t w and m initial The other is that which begins with a consonant followed by

either of the following l w j and r as in play tray and quick few The first consonant

here is called the initial while the second one post-initial Consonant clusters are up to

four Examples of three initial consonant clusters include split splɪt stream strim

square skwea Equally there are final consonant clusters which contain up to four

consonants two consonant cluster may include bump bᴧmp bent bɛnt bank bᴂŋk

belt bɛlt ask ᴂsk begged bɛgd and looked lʊkt among others There are two types of

final three consonant clusters final plus final plus post-final as in helps banks and bonds

and final plus post-final 1 Plus post-final 2 as in fifths (Roach 1983)

4131 EkeGusii and English phonotactics

Generally all the fourteen consonants in EkeGusii occupy the onset position while none

occupies the coda position because the language is a CV one as has already been

observed Equally all the vowels of the language take all the positions of a word initial

medial and final The same is not true of the English phonemes For example some

consonants in English do not occupy onset positions similarly others do not occupy coda

positions According to Cruttenden (2011) English does not exploit the syllable all possible

combinations of its phonemes For instance long vowels and diphthongs do not precede

final ŋ e ᴂ ʌ ɒ do not occur word finally and the types of consonant cluster permitted

are subject to constraints in both initial and final positions ŋ does not occur word

initially no combinations are possible with tint dʒ eth z r j w can occur in clusters only

as the non-initial element such initial element sequences as fs mh stl spw are not

allowed Finally only l may occur before non ndash syllabic m n h r j w do not occur in

word final positions and terminal sequences such as kf intp ɪ ʒbd are not used in the

language In the following sub sections 41311 and 41312 word initial word final

phoneme sequences and inflection suffix formation constraints of English are discussed

respectively

41311 English word initial phoneme sequences

Word initial consonant sound sequences in English vary from word to word There are

words with only a single consonant word initially while there with four consonants

Cruttenden (2011) observes that there are ten vowels in English which constitute

monosyllabic words as given in (34)

34) English vowels constituting monosyllabic words

vowel word

i e letter ltegt

ǝ a ɑ are

ͻ ɜ err

eɪ a letter ltagt

aɪ i letter ltigt

ǝʊ owe

ɪǝ ear

eǝ air

Adapted from Cruttenden (2011 201)

(34) shows that one vowel makes up an English word which is monosyllabic Cases of

vowels occurring word initially as syllables are common in English According to

Cruttenden (2011) all vowel sounds can occur word initially in English depending on the

word in question Thus some English words allow vowels in word initial position while

others do not

There are cases of consonant vowel (CV) in word initial positions with an exception of the

consonants ŋ and V All the other consonants generally occur before all vowels In

English also are cases of consonant consonant vowel (CCV) word initially In (35) there

are two consonant cluster patterns for English word initial positions as repeated from

Cruttenden (2011)

35) Two consonant cluster patterns for English

Cluster form Examples of words

P+l r j ply pray pure

t+ r j w try tube sweep

k+l r j w class crush cube

b+l r j blood breed beauty

d+ r j w dry dupe dwell

g+ l r j w glass grass

m+ j mew

n+ j new

l+ j lure leau

f+l r j flow fraud few

v+j view

θ+ r j w throw thief

s+l r j w p t k m n f v slow sir sue spree store skin smart

int+l r w m n shrewd

h + j hew

There are also cases of three consonant cluster patterns word initially in English as in (36)

repeated from Cruttenden (2011)

36)Three consonant cluster patterns for word initial position in English

Cluster form example of words

s+p+l r j splendid spring spying

s+t+ r j street skive

s+k+l r j w screw skew squad

As can be observed s is the essential first element of the CCC clusters the second

element being a voiceless stop the third element must either be l r j or w

41312 English word final phoneme sequences

There are cases of word final vowels in English Cruttenden (2011) observes that most of

the English vowels except e ᴂ ʌ andɒ occur word finally Concerning cases of final

vowels and consonant combinations Cruttenden observes that r h j w do not occur word

finally ʒ occurs finally only after the vowels i ɑ u and ei in words of recent French

origin like liege liʒ rouge ruʒ beige beiʒ ŋ occurs only after the vowels ɪ ᴂ ʌ

and ɒ There are also cases of VCC (vowel consonant consonant) combinations There

are a few mono-morphemic words of this kind including act adze axe corpse and lapse

The consonants r h j and w do not combine with other consonants in word final

positions in English (RP) g ŋ do not occupy final position in a final CC cluster θ is of

limited occurrence in this position

Cases of English final VCCC that is that of a vowel followed by a consonant cluster of

three do occur in English such as collapsed kɒlǝpst text tekst and prompt prɒmpt

These final CCC English clusters can be divided into two groups (i) those which involve a

combination of the two types of CC clusters that is m n ŋ l s plus C plus t d s z θ

These according to Cruttenden (2011) nearly all involve suffixes such as jumps cults lists

but there are monomorphic words such as mulct and calx (ii) Those which involve the

double application of t d s z θ the majority in this case involves suffixes such as

fifths fifθs products prɒdʌkts acts ᴂkts but there are two common monomorphic

words text tekst and next nekst (Cruttenden 2011) Cruttenden further observes that the

CCC clusters predominantly follow short vowels Eleven of the 49 CCC final clusters

occur after only one vowel (that is five after ɪ four after e one after ʌ and one after

Finally there are cases of VCCCC final word syllable The CCCC clusters occur only

rarely as a result of the suffixation to CCC clusters of t or s morpheme as in -mpts in

prompts exempts -mpst in glimpsed -lkts in mulcts -lpts in sculpts -lfθs

twelfths -ntθs thousandths Both of these word initial and word final phoneme sequences

indicate that there are cases of syllable complex margins in English

414 Comparison of EkeGusii toneme structure and English stress

This section presents a tonal description of EkeGusii noun as compared with English stress

The focus of the section is on the tonal patterns of EkeGusii noun in isolation and stress as

it characterizes the English noun

Tone has been defined differently by different phonologists According to de-Lacy (2007)

tone is the use of pitch in language to distinguish lexical or grammatical meaning

Languages that are characterized by these feature are known as tone languages Many

language of the world are tonal (Katamba 1993 and de Lacy 2007) Such languages

according to Katamba and de-Lacy have morphemes which are sometimes realized by

pitch changes that is using pitch differences to make phonemic contrasts In tone languages

therefore pitch can be used to distinguish word meaning or convey grammatical

distinctions It is in this respect that tone languages differ from non-tone (stress) language

like English where pitch does not have these functions

4141 EkeGusii tone structure

EkeGusii is a tone language (Bickmore 2007 Nash 2011 and Cammenga 2002) in

which pitch is used in the distinction of grammatical meaning more than lexical meaning

Examples of noun lexical contrasts based on tone are given in (37)

37) EkeGusii noun tone distinction

Word Tonal realization Gloss

(i) omogaaka oacute m ograve γ aacute agrave k a Old man106

omogaka oacute m ograve γ agrave k agrave aloe vera

(ii) omote oacute m ograve t eacute tree

omote oacute m ograve t egrave name of a person

(iii) esese ē s ē s ē dog

(iii) esese έsέsέ strain

This data shows that the distinction between the given words is as a result of contrastive

vowel length which according to Goldsmith (1990) is referred to as compensatory

lengthening and tone differences In compensatory lengthening vowels simultaneously

linked to several verb-slots are described as long and are at times phonemic that is

contrastive (Katamba 1993) The nouns omogaaka omoγaaka lsquoold manrsquo and omogaka

omoγaka lsquoaloe verarsquo are distinguished by the length of the first vowel of their roots as

illustrated by figures (18) and (19)

C V V C V

γ a k a -γaaka

Figure (18) EkeGusii lexical contrasts due to tone structure

C V C V

γ a k a -γaka

Figure (19) EkeGusii lexical contrasts due to tone structure Adapted from Katamba (1993)

Figure (18) shows that compensatory lengthening takes place when a single vowel is

doubly-linked with two verb slots in the underlying representation Thus a surfaces as

long [a] in omo-γaaka This is not the case in figure (19) where the vowel a is linked to a

single vowel slot thus surfacing as a short vowel Lexical contrast between the words in

figures (18) and (19) are based on tone distinctions This sub-section briefly describes

Ekegusii noun tone structure in which 41411 describes underlying versus surfaces tones

41412 Contour tones 41413 tone preservation and 41414 tone floating

41411 Underlying and surface tones in EkeGusii

Following Bickmore (1997) Pulleyblank (1986) Nash (2011) and others on Bantu

languages tone structure this study takes the view that that the underlying tonal distinction

in EkeGusii is one of high versus toneless that is low tones are underspecified

underlyingly only introduced at a later stage to the surface through insertion (Mwita

2012) This is in agreement with Kisseberth and Odden (2003) who observe that the surface

tone of the augment and the class prefix is normally low in Bantu languages Following

these observations therefore this study argues that Ekegusii has two basic surface tones

High (H) and Low (L)

Surface tones are marked by accent marks a transcription form used by Africanists (de-

Lacy 2007 230) High tone in this case is marked by an acute accent (acute) and represented

by H while a low tone is either unmarked or marked with (-) and represented by L This is

illustrated by figure (20) and is used in this study

EkeGusii tone realization word Gloss (i) o o kacutei m a obokima ugali

L L H L

(ii) om o tint acuteo k o racuteo omochokoro grand child

L L H L H Figure (20) EkeGusii tone marking

Adapted from de Lacy (2007 26)

Figure (20) above shows that the tone bearing units (morae or vowels) in class prefixes are

low (L) toned while the first tone bearing units in the roots are high (H) toned

41412 Contour tones in EkeGusiiAccording to Katamba (1993) Autosegmental Phonology Theory does not require a one-

to-one association of elements on different tiers (tonal tier segmental tier and CV tier)

Elements at any one tier may be linked one-to-many with elements in another tier The

following tonal examples in figure (21) of Mende language data (Leben 1978) repeated

from Katamba (1993 157) contain falling or rising tones Such tones are called contour

k כ bεlε mbu mba

H L H L L Hlsquowarrsquo lsquotrousersrsquo lsquoowlrsquo lsquoricersquo

Figure (21) Mende contour tones

Figure (21) shows that the Mende tonal contours are made when independent high and low

tones are simultaneously linked to a single vowel (Katamba 1993 157)

EkeGusii language like other Bantu languages Kuria (Mwita 2012) has a rising contour

(LH) when only one of the two consecutive vowels in a long syllable is marked for tone

that is it is high This is illustrated by figure (22)

Word tonal gloss word tonal gloss(Singular) realization (plural) realizationemoori e m ō oacute ri calfrsquo chimori c h i m ō oacute r i calves

L H LHomoonto om oacute oacute n t o Person abaanto ab ā aacute n t o Persons

L H L HFigure (22) Ekegusii LH contour toneAdapted from Katamba (1993)

This figure shows that the long syllable which starts with a low tone and ends in a high one

forms an LH contour The figure further indicates that this is when the root of the noun

starts with a vowel which is a copy of the prefix vowel However this is not the case when

the root starts with a consonant In such a case both consecutive vowels bear the same tone

marking and therefore the syllable is level that is it is pronounced with the same pitch

This is illustrated by figure (23)

Word Tone Gloss Word Tone Gloss (Singular) realization (plural) realization

obokokombe o o ndash koacute oacute m b e hoe amakombe ama ndash k oacute oacute m b e hoes

omogaaka omo ndash γ aacute aacute k a old man abagaaka abandashγ aacute aacute k a old men

H H HH

Figure (23) Ekegusii level tone

41413 Tone floating in EkeGusii

As has already been observed deletion of the vowels as in the examples given in data set

(45) above does not directly affect the tones which are associated with the vowels deleted

and as a result after the deletion of the vowels the tones simply remain on the tonal tiers

with no association with the segmental tiers This study like others such as Odenrsquos (2005)

is of the view that such an association creates floating tones Figure (24) illustrates this

observation

e g e n t o e k e e g e nt eke lsquothis thingrsquo

H L H L H H L HLH

Figure (24) EkeGusii floating tone

Figure (24) indicates that the high tone of the vowel o in egento lsquothingrsquo floats at the

surface It is this floating tone that is associated with the following vowel e (low toned or

high toned) resulting in a falling or rising tone (a contour tone) in this case being a falling

tone (HL)

41414 Stress in English

While EkeGusii is a tone language as has been discussed above English is a stress

language Verma and Krishnaswamy (1989) define stress as the intensity or prominence

with which a syllable is produced Thus in the production of a word or any syntactic

structure (sentence or phrase) there is always a syllable which is produced with

prominence It is this prominence that phonologists have come to refer to as stress

According to Laver (1994) if prominence is put on syllables on isolated words the

resulting stress is referred to as word stress Prominence given to words in sentences on the

other hand is known as sentence stress This study is interested in word or lexical stress in

particular noun lexical stress

Languages like English with syllables that differ in stress are stress languages This means

that these languages have more than one stress normally a loud or primary one which is

marked by a short raised stroke [] a medium or secondary one marked by a short lowered

stroke [sbquo] and an unstressed one which involves a non-prominent syllable containing no

pitch changes and has one of these vowels ɪ ʊ or ǝ (Laver 1994)

Depending on the number of syllables class of the noun and the nature of the word

whether compound or not a noun will be stressed differently Since every word has at least

one or more stressed syllables (Laver 1994) monosyllabic nouns have their only one

syllable stressed Equally bisyllabic nouns have their primary stress on the first syllable as

shown in (38)

38) English monosyllabic and bisyllabic noun stress

i) monosyllabic nouns ii) Bisyllabic words

maelign man pǝmɪt permit

strɪkt strict ekspͻt export

wik weak kɒntrʌkt

(38) shows that in all the words stress is placed on the first syllable of the given nouns

The following subsection gives a brief description of stress in the English noun

For nouns with three or more syllables (that is polysyllabic words) stress is determined by

the ending of the noun in question or generally the suffix (Laver 1994) Thus in nouns

which end in either ndasher or ndashly primary stress is placed on the first syllable just like in the

monosyllabics and bisyllabics above This is illustrated by (39)

39) Primary stress on polysyllabic nouns ending in -er or ndashly

ɒdǝlɪ orderly

maelignɪdʒǝ manager

415 Comparison of EkeGusii and English morphological structures

This sub-section focuses on the morphological processes that give insight into

morphological nativization of EkeGusii loan nouns from English It describes

morphological processes that explain word building processes in EkeGusii as compared to

English It describes the morphosyntactic classes of EkeGusii and in doing so the study

relies heavily on Cammengarsquos (2002) pioneering findings for EkeGusii This study

however unlike Cammengarsquos which is not anchored on any theory alludes to tenets of

Optimality Theory in its generalized descriptions

4151 Comparison of EkeGusii and English noun systems

According to Demuth (2000) noun classes in Bantu languages tend to be realized as

grammatical morphemes rather than independent lexical items These classes function as

part of a larger concordial agreement system where nominal modifiers pronominals and

verbs are all morphologically marked with the same noun class (gender) features Demuth

further observes that the classes are presently morphologically productive in most Bantu

languages and that semantically the classes have been reconstructed from Pro Bantu

Thus much of the semantics of current Bantu noun classes is no longer productive and in

some languages the number of classes has been reduced Demuth concludes that despite all

the given observations noun class systems especially morphologically are grammatically

productive in most Bantu languages and semantically productive to some degree Just as

Demuth (2000) notes EkeGusii nouns are characterized as grammatical morphemes and

function as part of a larger concordial agreement system

Comparatively most of the nouns in English unlike those in Bantu are realized as

independent lexical items This is cognizant of the fact that language morphological

typologies exist Haspelmath (2002) identifies three types of such languages typologies

isolating agglutinative and fusional He observes that some languages are close to ideal

types that is close to either completely isolating (such as Chinese and Vietnamese) or

agglutinative (such as Turkish) Most languages however are mixed types sharing features

of different given ideal types English and EkeGusii are mixed morphological typology languages What distinguishes

them however is the degree of fusion and or agglutination (index of synthesis) For

example grammatical relations are shown mainly by means of prepositions in English thus

resembling the patterns of isolating languages However the derivational and inflectional

morphologies of the same language are partly agglutinative and partly fusional EkeGusii

on the other hand like most Bantu languages like Kiswahili (Haspelmath 2002) is more

agglutinating than isolating Indeed in an index of synthesis given by Haspelmath

Kiswahili is ranked higher than English which therefore means that EkeGusii is more

synthetic or agglutinating than English In the following sub-sections the mophosyntactic

classes and prefixes of EkeGusii are described in relation to English morphology

4152 Morphosyntactic classes and prefixes in EkeGusii noun

Nouns in Bantu are classified into sets referred to as noun classes (Meinhoff 1899)

According to Welmers (1973) there are at least 22 of these noun classes in Pro-Bantu but

individual languages have less than the Pro Bantu number For example Kiswahili has 16

(Carsteins 1991 amp1993) Sesotho 15 (Demuth 2000) Kivonjo 16 (Pinker 1994) Aghem

12 (Aikhenvald 2000) EkeGusii 20 (Cammenga 2002 amp Ongarora 2009)

Morphosyntactically an Ekegusii noun consists of a prefix and a stem both of which

generally compulsory With an exception of a few classes the prefix carries number and

size features and has a (vowel) consonant vowel (V) CV syllable structure (Cammenga

2002) (40) represents EkeGusii noun class prefixes carrying number and size features as

repeated from Cammenga (2002199)

40) Morphosyntactic noun class prefixes in EkeGusii1 omo - 2 aβa-1a mo-1b Ǿ3 omo- 4 eme-5 eri- 6 ama-7 eke- 8 eβi-9 e- 10 chi-9a e-n- 10a chi-n-11 oro- 12 aka-14 oβo- 15 oko-16 a- 21 na

According to Givon (1972) Cammenga (2002) and Ongarora (2009) the choice of these

prefixes is determined by the semantics of the noun stems with which they occur In other

words the prefixes carry the gender number and size of the stems to which they are

appropriately (in terms of semantics) prefixed as illustrated by (41)

41) EkeGusii noun gender prefixationa) omonyaroka lsquogirlrsquo abanyaroka lsquogirlsrsquoomo ndash ɳaroka aβa- ɳaroka13PSG- girl 23PPL- girl lsquogirlrsquo lsquogirlsrsquo omogaaka aβa- gaaka 13PSG- lsquoold manrsquo 23P PL lsquoold menrsquo lsquoold manrsquo lsquoold menrsquo

b) ekerandi lsquogourdrsquo ebirandi lsquogourdsrsquo ekerandi eβi- randi 73PSG- lsquogourdrsquo 83PPL- lsquogourdrsquo lsquo gourdrsquo lsquogourdsrsquo eke- moni eβi- moni 73PSG- cat 83PPL cat lsquocatrsquo lsquocatsrsquo

Adapted from Ongarora (2006)In (41a) above the noun stem nyaroka lsquogirlrsquo denotes lsquohumanrsquo referent hence co-occur

with singular prefix omo- and a plural one aβa- while that in (41b) refers to an

inanimate referent randi lsquogourdrsquo and accordingly co-occur with the singular prefix eke

and the plural prefix eβi- Thus the mutual exclusivity of these prefixes stems from the

gender of the nouns (Givon 1972 amp Ongarora 2009) Table (2) shows EkeGusii prefixes

both in their singular and plural forms and their stems semantic determinants

Table (6) EkeGusii Prefixes and their Stems Semantic Determinants Prefix Noun stem semantics (meaning) determinants

Singular Plural

1 omo- 2 aβa- personal spiritual animate beings kinship terms ie God angles devils the spirits of the ancestors and kinship terms (human referents)

1b Oslash- 2aβa kinship terms (human referents)

3 omo- 4 eme- socioculturally relevant objects events or periods trees parts ofthe body (non-human referents)

5 eri-rii- 6 ama- various types of common nouns eg cultural or objects and location tools parts of the body fruits

5 eri- 6 ama- augmentative + or pejorative-7 eke- 8 eβi- inanimate mostly cultural objects some parts of the body some

animals some shrubs or plants language names

7 eke- 8 eβi- diminutive +or - pejorative

7 ke- no plural adverbs places names

9 e- 10 tinti- many names of animals socially or culturally relevant entities (place objects events) some concepts

9a e-n 10a tinti-n same as 9-10

11 oro- 10a tinti-n social cultural and some natural objects12 aka- 8 eβi- Diminutive

12 aka- 14 oβo- diminutive non-pejorative

14 oβo- ama- some body parts culturally relevant entities (objects places events activities) some crop names

14 oβo- no plural concepts

14 βo- no plural adverbs place names15 0ko- 6 ama- some body parts abstract nouns mostly referring to activities or

events conceived abstractically (usually without plural)15 ko- infinitive marker (together with word- final suffix ndasha expressing

activities or events16 a- [ase] lsquoplacersquo only no plural21 ɳa- no regular plural proper names of persons individual heads of

cattle and placesSource Cammenga (2002 201)

This table shows that occurrence of prefixes with noun stems roots are semantically

determined Thus the meaning of the stems to which the given prefix is attached plays a

major role in its choice In other words occurrence of a prefix is not haphazard and without

meaning

Nouns in English unlike in EkeGusii are not classified in terms of classes in the sense

described above In fact as can be observed in table (2) the class of a given noun in

EkeGusii like in other Bantu languages is determined by the prefix Prefixation in English

performs different functions such as marking opposite for example un- in lsquounlockrsquo

English according to Katamba (1993) is a language that is characterized by base word

morphology Base word morphology entails the study of the lowest indivisible level of a

morphological construction (Kiparsky amp Moahannan 1982) McCarthy (2002) observes

that an important feature of English which differentiates it from many other languages is

that it has a high proportion of complex words with an agglutinative morphology and an

equally large number of words with an isolated morphology Therefore as illustrated in

(42) English morphology is neither purely isolating nor purely synthetic

42) English morphemes

(a) (b)

read ndash able leg ndash ible

hear ndash ing audi ndash ence

en ndash large magnndashify

perform ndash ance rend ndash ition

In (a) the two morphemes affixed together are different respectively- free and bound while

those in (b) are both bound The difference as observed by McCarthy is attributable to the

history of English Most of the free morphemes in (a) belong to that part of the vocabulary

of English that has been inherited directly through the Germanic branch of the Indo-

European language family to which English belongs whereas the morphemes in (1b) have

been introduced or borrowed from Latin either directly or via French Again the words in

(a) are more common than those in (b) which reflects the fact that among the most widely

used words the Germanic element still predominates This leads to the conclusion that in

English there is a strong tendency for complex words to contain a free morpheme at their

core This is the argument this study is based on

Structurally most noun prefixes in EkeGusii unlike in English have a bi-morphemic form

Thus the prefix is divided into two elements an initial vowel sometimes referred to as an

augment or pre-prefix and the prefix per-se (Elwell 2005) The pre-prefix is described in

41221 the prefix in 41222 and the noun roots in 41223

41521 The pre-prefix or augment

The pre-prefix according to Elwell (2005) is a syllable added to the beginning of a word in

certain languages EkeGusii unlike English has such a syllable especially in noun number

and class marking prefixes and some monosyllabic words (in which case the augment is

just a single vowel) (43) gives the EkeGusii augment structure

43) EkeGusii augment structure

a)omote lsquotreersquo

i) o- mo- te lsquotreersquo

ii) e- mete lsquotreesrsquo

aug 3SG tree aug 4PL tree

b) eee lsquoyesrsquo

e-ee lsquoyesrsquo Adapted from Bosire and Machogu (2013)

The prefix omo- in (43ai) marks the class of the noun lsquotreersquo that is class three and

number that is singular while the prefix lsquoeme-lsquo in (43a ii) marks class four and plurality

The augment structures in (43) above is represented on syllable node in figure (25)

i) omot e σ

vc vc v σ σ σ

v c v c v

o m o t e

ii) e e e σ σ σ119

vv v v v v e e e Figure (25) EkeGusii prefix and pre-prefix syllable nodesAdapted from Katamba (1989)

The pre-prefixes in figure (25 i and ii) in each of the given words are made up of single

vowel syllables the vowel o- in (21 i) and e- in (25 ii) The output for class 1 affix is

lsquoomo-in figure (25 i)

This study is of the view that the vowels at the beginning of a prefix are tolerated because

without them the prefixes that result are those of classes 1b (oslash-) and 2b (oslash-) which carry

the meaning of kinship terms or sometimes when referring to nobody in particular (that is

neutrally) as illustrated by (44)

44) Ekegusii prefix of classes 1b (oslash-) and 2b (oslash-)

i) monto mo-nto lsquopersonrsquo banto βa-nto lsquopersonsrsquo

ii) tata tata lsquofatherrsquo batata βatata lsquofathersrsquo

These are described as follows

- mo- nto -βa- nto

1bOslash3PSG person 2bOslash3PPL person

lsquopersonrsquo lsquopersonsrsquo

-tata -βa- tata

1bOslash3PSG father 2b3 PPL fathers

lsquofatherrsquo lsquofathersrsquo

The nouns in (44) differ from those data (43) in that while those (44) lack arguments those

in (43) have In (44) where an argument lacks the nouns prefixed refer to nobody in

particular The form lsquotatarsquo for example is prefixless it demands neither a pre-prefix a

prefix nor both in the singular form (class1b) Of interest to note is the fact that its plural

form as can be observed is either that of noun classes (2) or (2b)

In commenting on augmentation and non-augmentation Cammenga (2002) observes that

while augmentation is the basic or regular state of affairs in EkeGusii morphology non-

augmentation which lacks an augment as in (43) above may be characterized as the

special case He further points out that generally both syntactic and semantic factors

determine whether or not a word may take an augment That is the presence or absence of

an augment is determined by lexical category membership and the semantics of the noun

stem as has already been observed In this respect therefore nominal prefixes in

morphosyntactic classes 1-8 and 11-15 may or may not be augmented while the prefixes in

classes 1b oslash- 9 (a) e-(n) 10 (a) tinti- (n) 16 a- and 21ɳa- are never augmented

This is the view taken in this study

The full EkeGusii prefix is generally made up of two parts an augment (pre prefix) which

is a vowel V and a prefix ndash proper which is made up of a consonant and a vowel CV-

Therefore an EkeGusii prefix takes the form V-CV which covers nominal prefixes in

classes (1-8) and (10-15) (Cammenga 2002)

Lexically class 5 prefix is regularly pre prefixed erindash in nouns of which the stem begins

with a vowel but non-pre-prefixed riindash in nouns of which the stem begins with a

consonant This is a case of phonologically determined allomorphy In other words it is the

sound at the beginning of the given noun which determines its pre prefixation or non pre

prefixation Otherwise the noun is one and the same thing (Cammenga 2002) (45)

adapted from Bosire and Machogu (2013) exemplifies this observation

45)EkeGusii class 5 prefixes (singular)

a) rii-toke lsquobananarsquo b) eri-iso lsquoeyersquorii-sosa lsquopumpkin leaves eri-ino lsquotoothrsquorii-raba lsquosoilrsquo eri-ogo lsquomedicinersquorii-mama lsquodumb personrsquo eri-eta lsquonamersquo

In data (45a) the nouns begin with a consonant and therefore do not allow augmentation

(45b) on the other hand begins with a vowel and therefore allows augmentation What

qualifies them as phonologically conditioned allomorphs is the fact that they take the same

prefix form in their plural that is ama- as in rii-toke ama-toke and eri-so ama-

iso Classes 9 e- and 16 a- prefixes consist of a vowel which may not be augmented A

brief general description of the regulations of the shape or quality of the augment structure

in EkeGusii language is presented as follows (46) gives EkeGusii noun prefixes adapted

from Cammenga (2002)

46) EkeGusii noun prefixes1 o-mo-1b Oslash2 a-βa-3 o-mo ndash 4 e-me-5 eri-rii-6 ama-7 eke-8 eβi-9 e-9a e-n10 tinti

10 a tinti-n11 oro-12 aka-14 oβo-15 oko-16 a-21 ɳa-

A number of observations about the pre- prefix shapes in data (46) can be made Firstly all

the prefixes with the form CV- allow pre prefixation except for those in classes 10 tinti-

10a tinti- n 21 ɳa- and allomorph rii- of class 5 discussed in (45) above

Secondly that the pre prefix is a copy of the prefix vowel except in the case of class 5

eri- 8 eβi- and 10 tinti- Thus the augment in noun prefixes may be accounted for by a

rule as in (7)

47) Noun prefix augmentation rule

The rule states that copy the vowel of prefix CV- to the left of the input such that any

non-low output vowel must be [+ mid] This according to Cammenga (2002) includes all

relevant prefixes and pre prefixes but appropriately excludes pre prefixation of all V-

shaped prefixes It ensures moreover that the [+high - mid] or high front prefix vowel of 5

ri- 8 (βi- and 10 (a) tinti-n is lowered to a [+high + mid] or upper mid front augment

vowel e

41522 The Prefix

The structure of EkeGusii prefix has been described by a number of studies (Ongarora

2009 Cammenga 2002 and Whiteley 1965) As has already been observed there are 20

of these classes as given in (48)

48) EkeGusii noun classesClass Examples Gloss 1 omo- [omoonto] person

[omwaana] child 2 aβa- [aβanto] persons

[aβaana] children 1bOslash- [Oslashβaaβa)] mother

[Oslashmaγokoro] grandmother [Oslashsokoro] grandfather

3 omo- [omotwe] head [omote] tree

4 eme- [emetwe] heads [emete] trees

5 rii- [riirok] foodrest eri- [eriiso)] eye 6 ama [amaγoko] footrests

[amaiso] eyes 7 eke- [ekerandi] gourd8 eβi- [eβirandi] gourds 9 e- [esese] dog

[eusi] thread 10 tinti- [tintisese] dogs

[tintiusi)] threads 9 a e-n-|e-n-βaata| - [embaata] duck

|e-n-raaγera| - [endaaγera] food10 a) tinti-n- |tinti-n-βaata| [tintimbaata] ducks

|tinti-n-raaγera| [tintindaaγera] foods 11 oro- [oroko)] firewood 10 a) tinti-n- [tintiŋko] pieces of firewood 12 aka- [akaana] small honey comb 8 eβi- [eβinana] small honey combs 12 aka- [akamoonto] small person 14 oβo- [oβomoonto] small persons

[oβosaatinta] manhood abstract nounno plural

15 oko- [ͻkͻ βͻͻkͻ] arm [okoγoro] leg

[oγoto] ear [okoruγa] cooking

6 ama- [amaoko] hands [amaγoro] legs [amato] ears

16 a- [ase] place [no plural]

21 ɳa- [ɳagera] blackie (cow proper name (no plural)(ɳaγeeŋke) name of a place (proper name no plural

The prefixes are underlinedSource Cammenga (2002)

Demuth (2002) observes that Bantu noun class systems can be characterized in two

typological terms first noun classes normally realized as grammatical morphemes and not

independent lexical items Second the class system that morphosyntactically function as

part of a large concordial agreement system where nominal modifiers pronominals and

verbs are all morphologically marked with the same noun class (gender feature)

This study like others in Bantu languages (Demuth 2002 Ongarora 2009 and Kayigema

2010) recognizes the fact that EkeGusii noun classes tend to be realized as grammatical

morphemes rather than independent lexical items In the following sub-section the noun

classes are presented in their various grammatical morpheme forms as identified in (48)

above In particular the descriptions in the sub-section focus on the rootbase morphemes

of the identified classes prefixes and pre- prefixes having been accounted for in this and

previous section

41523 EkeGusii noun class roots

It has already been observed in section (4121) that Bantu nouns are realized as

grammatical morphemes rather than independent grammatical items and that these

morphemes function as part of a large concordial agreement systems Therefore

description of EkeGusii noun involves among other processes the identification of the

various constituent grammatical morphemes including the root

Katamba (199341) observes that ldquohellipa root of a word is the irreducible core of that word

with absolutely nothing attached to it It is the part of a word that is always present

possibly with some modifications in the various manifestations of a lexemerdquo For example

lsquotalkrsquo in English is a root with the following word forms talk talk-s talk-ing and talk-ed

As can be seen the form lsquotalkrsquo cannot be reduced any further without losing its meaning

This is how this study views the roots which are described in the following subsections

according to their classes as identified in (48) above

41524 Class 1 omo- and 2 aba-

Nouns belonging to these classes are those within the meaning of personal spiritual and

animate beings kinship terms including Godgods angles and spirits as described in table

(2) above (49) gives examples of noun roots in classes (1) and (2)

49) EkeGusii noun roots in classes 1and 2

Noun surface underlying root gloss form form form

omoonto [omoonto] o- moonto -onto person

aug- 3psg- root

abanto [aaanto] a - a- anto] -anto persons

aug-3pl- root

omonyenyi [ͻmͻɳɛɳi] ͻ- mͻ- ɳɛɳi ɳeɳi butcher

aug- 3psg- root

abanyenyi [aaɳeɳɳ] a- a- ɳɛɳi ɳɛɳi - butchers

aug- 3pl- root

(49) shows that EkeGusii rootbase form is of either -CV or V-CV- form While the

form CV- obeys the Onset syllabic constraint the V-CV- form violates it even if it is

the realized form (Prince and Smolensky 1993)

41525 Classes 3 omo- and 4 eme-

Nouns in these classes are those within the meaning of events or periods trees and parts of

the body (50) shows noun roots within these classes

50) EkeGusii noun roots in classes 3 and 4

(i) Class 3

Noun surface Underlying Root Gloss

form form form

Omote [omote] omo-te [te] tree

Omotwe [Omotwe] o-mo-tue [-twe] head

Omogondo [omoγondo] o-mo-γoondo [-γondo] garden

ii) Class 4

emete [emete] eme-te [e] trees

emetwe [emetwe] eme-twe [twe] heads

emegondo [emeγondo] e-me-γoondo [-γoondo] gardens

The morphological behavior of these classes that is 3 and 4 is just like that of classes 1

and 2 in which case they are in their singular and plural forms respectively Cammenga

(2002) observes that the semantic motivation of classes 1 and 2 still appears to be

somewhat stronger in present day EkeGusii as compared to the other classes This indeed is

a correct observation because the nouns in classes 1 and 2 almost solely deal with animate

humans in singular and plural forms respectively However some animate human beings

such as the the physically and mentally challenged more often are taken to other classes

such as 7eke- as in in eke-rema lsquolame personrsquo and 8ebi- in ebi-rema lsquolame personsrsquo

5ri- as in ri-tiino lsquodumb personrsquo and 6ama- as in ama-tiino lsquodumb personsrsquo

41526 Classes 9 e- 10 chi- 9 (a) en- 10 (a) chin-

Nouns in these classes are those within the meaning of animals some events places and

objects (51) gives noun class roots of these classes

51) Noun class roots for classes 9 10 9 (a) and 10 (a)

Class 9 e-

Noun Surface underlying root glossform form form

esese [esese] e-sese [-sese] dog

etaaro [etaaro] e-taaro) [-taro] journey

ebuunda [eβuunda] e-βuunda [-βuunda] donkey

Class 10 chin-

chisese [tintisese] tinti-sese [-sese] dogs

chitaaro [tintitaaro] tinti-taaro [-taaro] journeys

chibuunda [tintiβuunda] tinti-βuunda] [-buunda] donkeys

The roots in these classes like those in Class 9 are similar in form However while those in

class 9 carry the singular form those in class 10 carry the plural meaning

Class 9a en-

embata [embata] e-n-βaata [βaata] duck

endangera [endagera] e-n-raaγera [raaγera] food

embori [embori] e-n-βoori [βoori] goat

Class 10a chin-

[tintimbaata] tinti-n-βaata [βaata] ducks

[endaaγera] chi-n-raaγera [raaγera] foods

[emboori] e-n-βoori [βoori] goat

These data show that while the outputs (surface forms) of the roots in classes 9 and 10 are

generally similar to their inputs (underlying forms) at least in structure and morphological

features those in classes 9a and 10a are not While the outputs of these classes (9a and 10a)

have voiced obstruents [b] [d] and [g] their input roots have [β] [r] and [γ] respectively

which are voiced fricatives This is due to nasal homorganicity and voicing dissimilation

explained earlier on What this means is that EkeGusii language does not have the voiced

obstruents They only emerge at the surface as prenasals due to phonological conditioning

41527 Classes 5 eri-rii- 6 ama- 7 eke- 8 ebi- 12 aka- and 14 obo-

These classes are marked by combination of corresponding singular and plural prefixes as

in (52)

52) EkeGusii classes 5 6 7812 and 14 prefixes

Singular plural

(a) 7 eke - 8 eβi-

12 akas - 8 eβi-

12 aka- 14 oβo-

(b) 5 eri- 6 ama-

These singular plural pairing of the given prefixes is explained as follows Firstly a word

from another class entering in any of the classes in (52a) gets the meaning of diminution

besides its basic meaning while when such a word is transferred to the classes in (52b) at

least the idea of augmentation is added to its basic meaning

Secondly as has already been observed prefixes in (52a) except that of class 12 function

as regular class prefixes This is in addition to marking diminution and augmentation just

described This according to Cammenga (2002) underlines the fact that it is the particular

combination of singular and corresponding plural prefix that constitutes some class and

determines its meaning Prefix class 12 aka- is the one exception since it expresses

diminutive meaning only This double function of prefixes in classes 5 6 7 8 12 and 14

entails that words belonging to a class marked by any of them cannot be transferred to the

class to which it already belongs regularly in order to express diminution or augmentation

Diminution according to Cammenga (2002) can be achieved through class transference

with pejorative connotation non-pejorative diminution through adjectival modification

and the expression of degrees of pejorativeness through a combination of these two means

with or without an added adverb are all exemplified in (53) as follows

53) EkeGusii diminution by prefixation

a) Diminution

o- mo- oNto o- mo- ke a- βa- anto a- ba- ke

aug- 1 - person aug- 1 - small aug- 2 - person aug- 2 - small

[omoonto ͻmͻkɛ] [aβaanto aβakɛ]

lsquoa small personrsquo lsquosmall personsrsquo

b) Pejorative or non-pejorative in increasing degrees

a- ka- mo- onto

aug- 12- 1 - person

[akamoonto]

lsquoa small personrsquo (pejorative or non-pejorative)

The plural ([oβomonto]) is always pejorative

a- ka- mo- onto a- ka- ke

aug-12- 1- person aug 12- small

[akamoonto aγake]

lsquoa very small personrsquo

c) Pejorative in increasing degrees

e- ke- mo -onto e- βi- mo- nto

aug- 7 - 1 person aug 8- 1 person

[ekemoonto] [eβimoonto]

lsquosmall personrsquo lsquosmall personsrsquo

e- ke- mo- oNto e- ke- ke

aug-7 ndash 1 ndash person aug ndash 7 ndash small

[ekemoonto eγeke]

lsquoa very small personrsquo ndash pejorative

Plural (eβimoonto eβike)

e- ke ndash moonto e-ke-ke -mono

aug-7 ndash 1 ndash person a-7 ndash small- very

[ekemoonto eγeke mono]

A very very small personrsquo (pejorative)

Plural [eβimonto eβike mono]

(54) shows examples of EkeGusii pre-prefixation or augmentation

54) EkeGusii augmentation

a) non ndash pejorative

o- moonto o- mo- nene a- ba- anto- a ndash ba- nene)

aug - 1- person aug-1 - big aug-2 -person ndash aug-2- big

[omoonto omonene] [abaanto abanene]

lsquoa big person lsquobig personsrsquo

(b) non ndash pejorative or more usually pejorative

rii ndash mo-Nto a- ma-mo-Nto

5 ndash 1 ndash person aug-6-1 persons

[riimoonto] [amamoonto]

rii- here means lsquobigrsquo just like ma ndash

c) abusive

rii- ke-mo-Nto a -ma-ke-mo-Nto

5- 7- 1 ndash person aug- 6- 7 ndash 1- person

[riikemoonto] [amakemoonto]

d) pejorative

rii-mo-oNto rii ndash nene a ndash ma- mo-oNto- a- manene

5 1 peson 5 ndash big aug- 6 ndash 1 person aug ndash big

[riimoonto riinene] [amamoonto amanene]

lsquovery big personrsquo lsquovery big personsrsquo

Adapted from Cammenga (2002 206-7)

In (54c) under augmentation the form [riikemoonto] is not acceptable in the view of the

researcher who s a native speaker The prefix stacking which brings in the prefix ke- of

class 7 does not seem to add any meaning to the whole structure of the word In fact the

class 7 prefix brings in a meaning of small so that the structure could mean rsquobig small

personrsquo which in view of this study does not sound correct

Classification and sub-classification of the English noun is different from that of EkeGusii

Classification of nouns in EkeGusii is determined by the prefix which is in turn controlled

by the semantics of the noun in question This is not the case in English

42 Phonological nativization of English nouns in EkeGusii

As mentioned in chapter 1 in this study each natural language has its own structural system

upon which the words are built An arguments based on universal grammar provides that

languages have certain basic properties that they tend to share However as Massamba

(1991) correctly observes it is quite unlikely that any two languages share exactly the same

structural forms (phonology and morphology) In other words in addition to the universal

grammar (UG) properties shared by all grammars each grammar has some peculiar

sequential constraints This section deals with objective two of the study that analyses the

phonological changes that English nouns borrowed into EkeGusii undergo during

nativization It focuses on the phonological features that EkeGusii and English grammars

do not share and how the English phonological system is adjusted so that it conforms to the

phonological constraints of EkeGusii grammar Analyses in this section and indeed the next

one (43) are carried out within the standard Optimality Theory (McCarthy amp Prince 1993

Prince amp Smolensky 19932004) and the data analyzed are those which were gathered in

the field (English nouns in EkeGusii) The loaned nouns are carefully and critically

examined for purposes of realizing their phonological changes and how the changes can be

accounted for within Optimality Theory perspectives

Nativization of English nouns borrowed into EkeGusii at the phonological level is basically

governed by the syllable structure of EkeGusii This is to say that a loaned noun normally

violates some constraint(s) of syllable well formedness in the target language in the process

of nativization In other words the loaned noun avoids the syllabic structure of the source

language in order for it to be accommodated in the target language It is this avoidance that

leads to conformity because the foreign structure is avoided at the expense of the native

one hence nativization For example many languages avoid cluster consonants and onsets

Other phonological features and processes besides the syllable structure also determine

nativization Phonological nativization in this study is analyzed under four broad headings

Segmental phonotactic prosodic and phonological processes Under segmental

nativization (421) the focus is on the consonants and the vowels phonotactic nativization

focuses on syllable structure in (422) prosodic nativization (423) focuses on tone while

phonological processes nativization (424) focuses on a number of processes

421 Segmental nativization

According to Sapir (1964) and Zivenge (2009) languages are loosely similar that is they

have slightly different inventories with some similarity In the same way there is some

loose similarity between English and EkeGusii languages A number of phonemes found in

the English noun are not found in EkeGusii phonological system However this does not

mean that there are no similarities at all between the two phonological systems In other

words in as much as there are English phonemes not found in EkeGusii phonological

structure there are some phonemes found in both languages (Anyona 2011) In order for

the English phonemes to be accommodated in the new EkeGusii phonological

environment two approaches were employed by the speakers substitution and deletion

This study considered the former because it was the most common approach the speakers

adopted This is discussed under nativization of vowel phonemes in 4211 and consonant

phoneme nativization in 4212 respectively

4211 Nativization of vowel segments

EkeGusii has a vowel system that is different from that of English in the same way

consonants of the two languages differ However the vowel difference between the two

languages is more pronounced as compared to that of consonants This is probably because

as Anyona (2011) points out English has more vowels as compared to EkeGusii language

Anyona points out that unlike EkeGusii language which has only pure vowels or

monophthongs sometimes characterized by length English has diphthongs and triphthongs

as well besides having more monophthong vowels comparatively Therefore there are

many English vowels that are not found in EkeGusii phonology Thus most of the lsquoexcessrsquo

vowels from English are collapsed into the few EkeGusii vowels In other words while

English has twenty- five vowels (Cruntenden 2011 OrsquoConnor 2011 and Roach 1983

among others) EkeGusii has fourteen as has already been observed in this study Sub-

section 42111 analyzes nativization of English pure vowels 42112 with English

diphthongs while 42113 analyzes thriphtongs

42111 Nativization of English pure (monophthong) vowels

These are those vowels which when produced the tongue remains constant in that it does

not glide This sub-section shows how these vowels are integrated into EkeGusii

phonology

Nativization of English ɪ into EkeGusii [i] ( ɪrarr[i])

The English vowel ɪ shares almost similar features with the EkeGusii vowel [i] They are both

[+HIGH -ROUND AND ndashBACK] However while the English ɪ is [+LAX] EkeGusii [i] is

[ndashLAX] which explains why they are acoustically different as illustrated by tables (3) and (4) in

section (411) above The English vowel ɪ therefore was realized as EkeGusii [i] as in (55)

This is in addition to other phonological changes The substituted vowels are in bold

(55) Nativization of English ɪ to EkeGusii [i]English word Pronunciation EkeGusii nativized form pronunciationChristmas krɪmǝs ekirisimasi [ekirisimasi]guitar gɪtɑ egiita [egiita]kitchen kɪtintǝn ekicheni [ekitintɛnicabbage kǝbɪdʒ ekabichi [ekaitinti]The realization in (55) is expected because the two vowels are closely related in terms of

phonological features as has already been observed They are [+FRONT +HIGH AND ndash

ROUND] differing only in [LAXNESS] while [ɪ] is completely laxed [i] on the other

hand is slightly more tensed though not as much as the long [i] (OrsquoConnor 1967

Cruttenden 2011) In fact the vowel [i] is present in both EkeGusii and English (Anyona

2011 Cammenga 2002) differing only in their degree of tenseness during production

This is further supported by the acoustic differences between the vowels To demonstrate how the English pure vowel ɪ in (55) was substituted for by the EkeGusii

vowel [i] the word kirisimasi kirisimasi lsquochrismasrsquo is presented in figure (26) Oslashk r ɪ s m ǝ s Oslash English k Oslash r Oslash s Oslashm Oslash s EkeGusii

k r i s m a s Phonemic substitution

e k i r i s i m a s i (Nativization Borrowing complete)Figure (26) Substitution of the English ɪ for EkeGusii [i] Structural presentation adapted from Gussenhoven amp Jacobs (2011)

Figure (26) shows that the English front short high and unrounded vowel ɪ is

substituted for the EkeGusii front high tense and unrounded [i] vowel That is while [i] is

tense ɪ is lax This phenomenon is explained by the fact that the EkeGusii phonological

system does not have the lax front high vowel ɪ but both vowels that is ([ɪ] and [i])

share many common features they are [+High] [-Back] and [-Round] (Cruttenden 2011

Roach 1983 OrsquoConnor 1967) This is in agreement with Kang (2011) who argues that a

foreign input containing a segment absent in the target language necessitates the

replacement of the foreign segment by the closest sound in the target language Kang gives

the example of the adaptation of the French high front rounded vowel [y] as [u] (which

has the rounding and high qualities) in White Hmong as discussed by Golston and Yang

(2001)

In essence the realization of the English [ɪ] as EkeGusii [i] involves phonetic featural

changes Thus EkeGusii prefers tense vowels to lax ones The occurrence in figure (27) is

against Optimality Theoryrsquos markedness constraint TENSE (V) which prefers lax vowels

to tense ones (McCarthy 2007) The realization therefore bans lax vowels- LAX (V) The

different realizations of the vowels in the input (English) and output (EkeGusii) imply the

faithfulness constraint IDENT IO (FEATURE) which demands that an input feature must

also be in the output no change (Kager 1999) Since change is allowed at the expense of

having tense vowels the markedness constraint dominates the faithfulness one resulting to

the ranking argument LAX (V) gtgt IDENT IO (FEATURE)

OT differentiates languages on the basis of hierarchical ranking of universal constraints and

not on language particular recursive rules of early generative theories The ranking and re-

ranking of constraints in this study used the Tesar and Smolensky (1993) algorithm model

which provides that given surface forms of the borrowing language (in this case EkeGusii

nativized forms from English) and a set of universal constraints it is possible to discover

the correct ranking of the target language In this model it is assumed that an input that is

the form from which the output derives is provided (the English forms in the case of this

study) and that the output is the phonologically structured representation and not a raw

phonetic form (in this case the EkeGusii nativized forms from English collected from the

field) Given that the initial state of the algorithm is one in which all constraints are

unranked with respect to one another that is all are undominated the algorithm employs

the principle of constraint demotion in ranking and reranking of the universal constraints in

a language specific manner

Using the constraints given above English and EkeGusii realizations of the word

lsquoChristmasrsquo krɪmǝs and ekirisimasi [ekirisimasi] respectively are analyzed in tableaux

(1) and (2) respectively English realizationInput krɪsmǝs

This realization and indeed all the realizations in which vowels of the target language

(EkeGusii) are substituted for those of the source language (English) will rank the

constraints given above as follows

IDENT IO (FEATUREPLACE) V gtgt LAX (V) which means that IDENT IO

(FEATURE) VOWEL is ranked higher and therefore dominates LAX (V) Thus IDENT

IO (FEATURE) VOWEL plays an important role in determining the optimal candidate in

English In all the ranking arguments and how optimal candidates (winning candidates) in

this study are established and illustrated violation tableaux are used (McCarthy 2007

2008) This is because the goal of this study is to establish or select the optimal candidate

in the given constraint ranking Following this therefore the English realization above uses

violation tableau (1) to establish the ranking argument and demonstrate how the optimal

candidate competitively emerges

Input krɪsmǝs IDENT IO (FEATURE) (V) LAX (V)

a [krɪsmǝs]

b [kirisimasi]

Tableau (41) English realization of the input krɪsmǝsThe winning candidate here is (a) It satisfies the faithfulness constraint IDENT IO

(FEATURE) (V) which is highly ranked in English Its violation of LAX (V) is not fatal

since English allows it Candidate (b) loses because it violate the highly ranked constraint

in the language that is IDENT IO (FEATURE) (V) This is comparable to EkeGusii

output of the same word in tableau (2) belowInput kirisimasi lsquoChristmasrsquoThis realization re-ranks the constraints as follows LAX (V) gtgt IDENT IO (FEATURE)

(V) Thus it reverses the ranking The realization is analyzed in tableau (2)

Input kirisimasi LAX (V) IDENT IO (FEATURE) (V)

a [krɪsmǝs]

b[kirisimasi]

Tableau (42) EkeGusii realization of the Input kirisimasi

In this tableau the optimal candidate is (b) This is irrespective of the fact that the

candidate violates the faithfulness constraint IDENT IO (FEATURE) (V) as illustrated by

the tableau The faithfulness constraint is dominated by the markedness one in EkeGusii

unlike in English On the contrary candidate (a) loses because it violates a highly ranked

constraint LAX (V) which disallows lax vowels This in Optimality theory terms is a

fatal violation

The realization of the English ɪ as [i] in English nouns borrowed into EkeGusii is not a

peculiarly EkeGusii phenomenon English loans in White Hmong language spoken in in

Southern Chaina behave the same as illustrated by (56) below

56) White Hmong nativization of English lax vowel ɪEnglish word pronunciation Hmong realizationMcKinley mǝkkɪnli [mekiŋli]Mitsubishi mɪtsǝbiinti [miintimbiinti]Adapted from Goldstone and Yang (2001)(56) shows that the English lax vowel ɪ (bold) is realized as White Hmong tense [i] (bold)

Golstonersquos and Yangrsquos conclusion that short vowels not found in Hmong are borrowed into

the language as the vowel that is closest to them in terms of features such as height

rounding and backness seems to be the case in this study Indeed all the English lax

vowels entering EkeGusii were generally tensed as shown by data set (55) above [advise

on how to handle delete or leave]

Nativization of English aelig into EkeGusii [a] ( aeligrarr[a])

The vowels [a] and [aelig] differ only in one respect while the English aelig is lax EkeGusii [a]

is tensed accoustically They are the same in all other aspects they are [front non-

rounded low]

The English vowel aelig is realized as [a] by EkeGusii speakers as in (57)

57) Nativization of English aelig to EkeGusii [a]English noun pronunciation EkeGusii nativized form pronunciationtaxi taeligksi etagisi [etaγisi]glass glᴂs ekerasi [ekerasi]tank tᴂŋk etanki [etaŋgi] bathroom bᴂethrum ebaturumu [eaturumu] In (57) the English vowel aelig is realized as [a] in EkeGusii These two vowels share

phonetic and phonological features as has already been shown This explains why the

speakers substitute one for the other The only difference which is responsible for their

phonemic status is the phonetic feature [TENSE] [a] is [+TENSE] while aelig is [ndashTENSE]There are three possible explanations for the occurrence in (57) The first one is phonetic as

explained by Yip (2002) This provides that since aelig and [a] both have a lowered jaw in

their production the speakers find [a] a better perceptual match for the English aelig since

the muscles of the speakers are used to this production The second explanation which is

equally phonetic and closely related to the first one is acoustic EkeGusii unlike English

does not allow lax vowels thus English aelig which is lax is realized as [a] which is tensed

in EkeGusii (see section411) above The third explanation which is visual is that of

orthographic influence In this case as Peperkamp (2006) observes adaptations reflect the

way native speakers are used to reading of foreign graphemes According to Peperkamp

French children learn to pronounce English graphemes as their native sounds This is

illustrated in (58) as adapted from Peperkamp (2006)

58) Realization of English graphemes by French childrenEnglish Grapheme French Realization (Pronunciation) Example of word

ltugt œ butltoogt ltugt book

As a result of this Peperkamp observes that French adult speakers are likely to base their

adaptations of English words on these between language grapheme to phoneme

correspondence Both the phonetic and perceptual explanations seem to influence the realization of the

English vowel aelig as [a] in EkeGusii besides closeness in terms of phonological features

discussed in sub section 411 above (57) above indeed shows that all the noun loans the

vowel aelig is realized as [a] Optimality Theory account of this realization is the same as

that discussed in section 42111 above

Nativization of English ᴧ into EkeGusii [a] ( ᴧrarr[a])

The vowelsᴧ and [a] are characterized by similar feature values They are both [-BACK]

and [-HIGH] But while the English ᴧ which is absent in EkeGusii phonology is

[+LAX] EkeGusii [a] is [ndashTENSE] It is the phonetic similarity and difference that makes

it possible for the realizations witnessed in (59)

59) Nativization of English ᴧ to EkeGusii [a]

English noun pronunciation EkeGusii nativized form pronunciationcut kᴧt ekati [ekati]brush brʌint eburasi [eurasi]cupboard kʌbǝd ekabati [ekaati] pump pʌmp epambu [epambu]In (59) the English vowel ᴧ is realized as [a] by EkeGusii speakers In fact this is one of

the vowels which did not provide much pronunciation challenge to the speakers This is

perhaps because the two sounds are produced by almost the same part of the tongue and

their degree of tongue height is almost similar as illustrated by chart (6)

Front central backHigh Mid ʌLow a Chart (6) English ʌ and EkeGusii aChart (6) shows that both sounds that is ʌ and a are [+front] [+low] and [-rond]

Thus the sounds share more phonetic features values than they differ

Nativization of English ɜ and ǝ into EkeGusii [a] ( ɜ ǝ rarr[a])

The vowels English ɜ and ǝ are characterized by the phonetic feature values [+tense

-round -low] -front] while the EkeGusii vowel [a] is characterized by [+tense -round

+low front] Both the English and EkeGusii vowels share two features ([+tense ndashround])

which perhaps together with perceptual closeness determines the substitutions that occur

as illustrated by (60) and (61)

(60) Nativization of English ɜ and to EkeGusii [a]English noun pronunciation EkeGusii nativized form pronunciationskirt skɜt esikati [esikati]shirt intɜt esati [esati]breakfast brekfɜst burekibasiti [urekiasitinurse nɜs omonasi [omonasi]

(61) Nativization of English ǝ and to EkeGusii [a]English noun pronunciation EkeGusii nativized form pronunciationfather fɑethǝ omobaata [omoaata] christmas krɪmǝs ekirisimasi [ekirisimasi]pastor pɑstǝ omobasita [omoasita]computer kǝmpjutǝ ekombiuta [ekompjuta]

As (60) and (61) show the English ɜ and ǝ are realized as EkeGusii [a] (in bold) This is

a common phenomenon in loan word nativization For example Dholuo a Nilotic language

nativizes the two central English vowels which are absent in its phonology to [a] (Owino

2003) Bantu languages like Tonga and Kalanga spoken in Zimbabwe and Botswana

respectively (Zivenge 2009 Chebanne and Phili 2015) like EkeGusii also substitute the

English ɜ and ǝ for [a] Language family does not seem to determine the substitution

rather the absence of the vowels in the borrowing languages One feature value that the

English vowels do not share with the vowel it is substituted for in EkeGusii and the other

languages that is [a] is [+ CENTRAL] This is a marked feature value because many

African languages avoid it at the expense of either [FRONT] Theoretically therefore the

realizations of a instead of ɜ and ǝ in (60) and (61) respectively presuppose the

markedness constraint CENTRL (V) which prohibits central vowels but the change of

the feature values violates the faithfulness constraint IDENT IO (F) Tableaux (3) and (4)

ranks and re-ranks the constraints of English and EkeGusii realization of the English word

shirt intɜt for exampleEnglish input intɜt

Input intɜt IDENT IO (F) V CENTRAL (V)

a [intɜt]

b esati

Tableau (43) English realization of the input intɜt 143

EkeGusii input [esati]

Input esati CENTRAL (V) IDENT IO (FEATURE) (V)

a [intɜt]

b [esati]

Tableau (44) EkeGusii realization of the input esati

Re-ranking of the given constraints yields different outputs When the ranking is such that

the markedness constraint CENTRAL (V) dominates the faithfulness one IDENT IO (F)

(V) that is CENTRAL (V) gtgt IDENT IO (F) (V) as in tableau (4) EkeGusii output

results The opposite is true when the faithfulness constraint dominates the markedness

constraint as in tableau (3) Thus English tolerates the given markedness constraint as

compared to EkeGusii and other African languages

The interpretation of tableau (3) for the English output is that candidate (a) is the output

because it obeys the faithfulness constraint which is ranked higher in English as compared

to EkeGusii The markedness constraint on the other hand dominates the faithfulness

constraint in tableau (3) to enable candidate (b) to be the output

Nativization of English ɑ into EkeGusii [a] ( ɑ rarr[a aa])

These vowels share the feature values [+low +tense and -round] They differ in that while

the English ɑ is [-FRONT] EkeGusii [a] is [+FRONT] The choice of [a] as a substitute

therefore is expected because the two vowels share many feature values than they differ

The substitution of ɑ which is [-FRONT] for [a] which is [+FRONT] presupposes the

markedness constraint BACK (V) which prohibits back vowels they especially [+LOW]

ones are marked (Kager 1999) Thus as (62) indicates all cases of the English ɑ coming

into EkeGusii phonology were realized as either [a] or [aa] which is [+FRONT]

(62) Nativization of English ɑ to EkeGusii [aa]English noun pronunciation EkeGusii nativized form pronunciationcar kɑ ekaa [ekaa]card kɑd ekati [ekaati]glass glɑs ekerasi [kerasi]garage gaeligrɑʒ egarachi [γaratinti]

In (62) the open low back tense English vowel ɑ is realized as EkeGusii [aaa] This is

because EkeGusii does not have the English vowel ɑ in its phonological inventory and

most importantly ɑ a low back vowel is marked Closer orthographic perception also

plays a role The markedness feature which is the main determinant of the substitution

presupposes the markedness constraint BACK (V) which bans back vowels This

constraint in turn means that the faithfulness constraint which demands that input and

output features be the same (IDENT IO (F)) is violated Thus the outputs of the English

input ɑ in English and EkeGusii is determined by re-ranking of these constraints as

analyzed by tableaux (5) and (6) for the English word glass glɑ sEkeGusii input ekerasi

Input ekerasi BACK (V) IDENT IO (F) V

a [glɑs]

b [ekerasi]

Tableau (45) EkeGusii realization of the input ekerasi English input glɑs

Input glɑs IDENT IO (F) V BACK (V)

a [glɑs]

b [ekerasi]

Tableau (46) English realization of the input intɜt

In tableau (5) candidate (b) is the output because it obeys the markedness constraint

BACK (V) which bans back vowels Its violation of the faithfulness constraint is

inconsequential because the constraint is lowly ranked in EkeGusii The reranking of the

constraints leads to the analysis in tableau (6) English ranks the faithfulness constraint

higher than the markedness constraint which is why it tolerates the marked feature

BACK which is avoided by EkeGusii

This realization is not peculiar to English nouns borrowed into EkeGusii Hmong Golston

and Yang (2001) Dholuo Owino (2003) Tonga (Zivenge 2009) and Kalanga Chebanne

and Phili (2015) among others behave the same way For example in Dholuo Owino

(2003) just like in EkeGusii the vowel ɑ is realized as Dholuo [a] as in (63)

63) Nativization of English ɑ to Dholuo [a]English noun pronunciation Dholuo nativized form pronunciationgarage gaeligrɑdʒ garach [garatint]glass glɑs gilas [gilas]card kɑd kadi [kadi] Source Owino (2003) This realization further confirms the fact that [+LOW] [+BACK] vowels are marked and

therefore absent in most languages of the world because they are not easy to learn and

produce

Nativization of English ɒ into EkeGusii [ͻ] ( ɒ rarr[ͻ])

The vowels [ɒ] and [ͻ] are characterized by the value features [+back +round] But while

the English ɒ is [+LOW] EkeGusii [ͻ]) is [-LOW] Thus this is the feature which

determines the substitution of the [+LOW] vowel for the [-LOW] one As observed by

Kager (1999) [+LOW +BACK] vowels are marked and therefore avoided by most

languages English ɒ is avoided in EkeGusii as in (64)

(64) Nativization of English ɒ to EkeGusii [ͻ] English noun pronunciation EkeGusii nativized form pronunciationCotton cɒtn ekotini [ɛkͻtoni]Box bɒks epogisi [ɛpͻγisi]Bolt bɒlt eboriti [ɛͻriti]In (64) the English short back rounded English vowel ɒ is realized as EkeGusii [ͻ]

Tableau analysis of this realization is the same as those of the realization in (62) above

because it is the same markedness constraint involved in both cases that is BACK V

Nativization of English ɪ into EkeGusii [e] or [ɛ ] ( ɪrarr[e] or [ɛ])

These vowels are characterized by the following phonetic feature values [-BACK

-ROUND -TENSE] The only feature which distinguishes the two vowels is [high] while

the English ɪ is [+HIGH] EkeGusii [e] is [-HIGH] This is perhaps one of the reasons

behind the realization of the English ɪ as EkeGusii [e] or [ɛ] as in (65)

(65) Nativization of English ɪ to EkeGusii [e]English noun pronunciation EkeGusii nativized form pronunciationcollege kɒlɪdʒ ekorechi [ɛkͻrɛtintisenate sɪneɪt eseneti [seneti]elephant elɪfǝnt erebanti [ɛrɛanti]

In (65) the English vowel ɪ is realized as EkeGusii [e] or [ɛ] The choice between [e] and

[ɛ] is determined by vowel harmony discussed in section 41111 above This realization

can be given two explanations The first explanation is that of orthographic influence

which is perceptual or orthographic in nature (Owino 2003 Peperkamp 2006) as

discussed in section 41112 above The orthographic system of the vowels and not the

feature values of the vowels dictate the pronunciation of the English ɪ as EkeGusii [e] or

[ɛ] For example in college the letter ltegt influences the realization of [ɛ] which is closely

related to [e] in terms of phonetic features and not the English ɪ which is neither in the

EkeGusii orthography nor closer featurally to the [ɛ] The second explanation is phonetic It

has already been observed in this section that the vowels are more similar phonetically than

they differ they differ only in terms of [height] while [ɪ] is [+high] [e] and [ɛ] are [-high]

Markedness has it that high vowels are more marked as compared to low vowels

(Trubetzkoy 1969) thus [ɪ] is more marked and therefore less natural than [ɛ][e]

Therefore it is easier to produce [e] [ɛ] as compared to [[ɪ]

The realization in (65) like that of other vowels discussed so far show a change of feature

values between the input and output forms of the English vowel ɪ Thus in OT theoretic

terms faithfulness constraints such as IDENT IO (F) V and markedness constraint

ASSIM (F) which prohibits assimilation of features in a given domain are presupposed

Thus while English demands that the vowel in the input must be preserved in the output

EkeGusii demands that vowels be assimilated This results in different ranking of the

constraints as demonstrated by analyses of the English and EkeGusii outputs of the word

senate sɪneɪt for example in tableaux (7) and (8) below respectivelyEnglish realizationInput sɪneɪt lsquosenatersquoConstraints ranking IDENT IO (F) V gtgt ASSIM (F)

Input sɪneɪt ASSIM (F) IDENT IO (F) V

a [sɪneɪt]

b [eseneti]

Tableau (47) English realization of the input sɪneɪtEkeGusii realizationInput esenetiThis realization is presented in tableau (8)Constraint ranking IDENT IO (F) gtgtASSIM (F)

Input eseneti IDENT IO (V) ASSIM (F)

a [sɪneɪt]

b [eseneti]

Tableau (48) EkeGusii realization of the input eseneti

In tableau (7) the optimal candidate is (a) because it does not violate the constraint

ASSIM (F) which is the highest ranked while in tableau (8) candidate [b] wins because it

satisfies the constraint IDENT IO (V) which is banned in EkeGusii

Other languages for example Dholuo Owino (2003) also sometimes nativize the English

vowel ɪ to [e] as demonstrated by (66)

66) Dholuo nativization of English ɪ to [e] English noun pronunciation Dholuo nativized form pronunciationmission mɪintn misen [misen]television telɪvɪintn telefison [telefison]elephant elɪfǝnt elefant [elefant]

In (66) the English vowel ɪ is realized as [e] in Dholuo just like in EkeGusii as shown in

(65) Just like in EkeGusii this vowel is not present in Dholuo phonology(Owino 2003)So far under the section of English pure vowel nativization it has been realized that in

EkeGusii loaned words from English the central vowels ɜ ǝ ʌ are substituted for

EkeGusii low front vowels [a] as illustrated in figure (27)

ǝ [a]

ʌFigure (27) EkeGusii nativization of the English central vowels SourceBright (1970 123)

These realizations according to Owino (2003) can be attributed to phonetic factors Owino

observes that on account of restricted physiological space associated with the lower region

of the oral cavity it can be assumed that the articulatory and auditory properties of the low

vowels occurring in English are minimally differentiated At the same time Dholuo

operates on a single low vowel a This vowel can be considered a natural rendering of the

central vowels found in English This is in fact what characterized the English central

vowels and indeed most other vowels upon coming into EkeGusii phonology In other

words some of the English vowels entering EkeGusii like those entering Dholuo are

collapsed into the EkeGusii [a] and the other few EkeGusii vowels This of course is

dependent on the phonological closeness (in terms of features) between the target and the

source language and sometimes the hardness with which the incoming vowel is produced

that is its markedness status The integration of English vowels into the vowels of EkeGusii

is further illustrated by (82)

67) English vowel realization in EkeGusiiEnglish vowel EkeGusii realization

i ɪ [i]ʌaelig ǝ ɑ [e ɛ] e aelig ɜ [e]u ʊ [u]ͻ ɒ [o ͻ]

Source Anyona (2011)(67) shows that all the English vowels are collapsed into the seven EkeGusii vowels This

is in agreement with Bright (1970) who observes that African languages collapse English

vowels into those present in their phonologies as in (68)

68) English pure vowels against their approximate African vowel phonemesEnglish vowel Approximate African languages vowel

i ɪ [i]ʌ ǝ ɜ ɑ [a]e aelig [e]u ʊ [u]ͻ ǝʊ [o]

Adapted from Bright (1970)Cases of vowels of source languages being substituted for those of the target languages as

in the case of this study are common (Golstone ampYang 2001 Owino 2003 Zivenge

2009 Hussain 2011 Kang 2011and Chebanne amp Phili 2015) All these studies like the

present one point to the fact that vowels and indeed sounds of the source language change

to or are substituted for those of the target language when they are not present in in the

phonologies of the target languages For example as (56) above attests the English

vowel ɪ is substituted for the White Hmong [i] just like in EkeGusii The difference

between White Hmong and EkeGusii studies being that while Hmong is a Miao-Yiao

language which is isolating and largely monosyllabic spoken in Southern Chaina

(Golstone and Yang 2001) EkeGusii on the other hand is a Bantu language which is fairly

polysyllabic and agglutinative spoken in Kenya

The fact that the two languages share the given phonological phenomenon is illuminating

It means that the shared feature would be regarded as a universal tendency This

observation is further supported by the fact that findings in the other studies mentioned

above point to the same direction irrespective of the fact that some are accounted for by

different theoretical perspectives while others are not anchored on any theoretical

perspectives Golstone amp Yang (2001) and this study are anchored on Optimality Theory a

constraint based generative theory while Owino (2003) and Zivenge (2009) are accounted

for within rule based generative theories Hussain (2011) and Chebanne amp Phili (2015) on

the other hand are not anchored on any theoretical framework yet the results of all the

studies are the same The findings of all these studies point to the universal nature of the

given phonological occurrence that is the substitution of foreign language vowels not

present in the target language for those present in the target language

42112 Nativization of English diphthongs

A diphthong according to Roach (1983) is a vowel containing two vowels pronounced as

one gliding from one to the next in rapid succession EkeGusii unlike English does not

have diphthongs EkeGusii vowels which follow one another in a word like those in other

Bantu languages are not realized as single units forming syllable nuclei in English Thus

such vowels in Bantu get realized as two distinct vowels belonging to two successive

syllables instead of forming the nucleus of a single syllable as is the case in English

(Chebanne and Phili 2015)

Many cases of diphthong nativization are realized as single vowels in the borrowed words

as observed under the section of phonological processes nativization in this study The

common occurrence is that English diphthongs are either substituted for either by a single

phoneme vowel (monophthongization) or lose its second element and lengthen the first

element This in Optimality theory suggests the following constraints IDENT IO (F) a

faithfulness constraint which demands that features of an input segment must be preserved

in the output no feature change and COMPLEX V a markedness constraint which bans

complex vowels and MAX IO another faithfulness constraint which demands that input

segments must have output correspondents deletion of segments is disallowed This

subsection shows how the English diphthongs were realized in EkeGusii and how the

realizations are accounted for within Optimality Theory

Nativization of English ǝʊ into EkeGusii [o] ( ǝʊrarr[o])

The diphthong ǝʊ is made up of two pure vowels the first of which being mid central and

non-rounded while the second is high back and rounded The EkeGusii [o] on the other

hand is a monophthong with the feature values [+ MID] and [+ ROUND] It has one

feature from each of the vowels of the English diphthong it substitutes [+MID] from [ǝ]

and [+round] from [ʊ] This probably explains why the diphthong is substituted for the

monophthong in EkeGusii realizations as in (69) below

69) English əʊ nativized to EkeGusii [o]English noun pronunciation EkeGusii nativized form pronunciationLocation lǝʊkeɪintn erookeseni [ero keseni]Sofa sǝʊfǝ esooba [esoa] Cocoa kɒkǝʊ ekooko [ekoko]Pawpaw pǝʊpǝʊ ripoopo [ripopo]Radio reɪdiǝʊ ereetio [ereetjo]

In (69) the English diphthong ǝʊ is realized as the EkeGusii back mid rounded vowel

[o] This is achieved through the process of vowel coalescence which is a common process

affecting vowel nativization (Zivenge 2009 Owino 2003 Chebanne amp Phili 2015)

Orthographic influence according to Golstone amp Yang (2001) and Hussain (2011) is

another reason behind this realization In other words orthography makes speakers focus

on the grapheme rather than the actual sound The change noted here is that of coalescence

which in essence leads to the substitution of [o] for ǝʊ in EkeGusii This occurrence

presupposes the OT constraints given above in the given ranking COMPLEX V gtgt

MAX OI IDENT IO (F) This is because the realization of the diphthong which leads to

monophthongization satisfies the constraint COMPLEX V (which is highly ranked in

EkeGusii) at the expense of violating the constraints MAX OI and IDENT IO (F) (both of

which relatively lowly ranked in the language) To account for realizations in (84) the

EkeGusii nativized word form [erokeseni] lsquolocationrsquo for example is presented in tableau

EkeGusii input ero keseni Constraint ranking COMPLEX V gtgt MAX OI IDENT IO

Input erokeseni COMPLEX V MAX OI IDENT IO (F)

a [lǝʊkeɪintn]

b [erokeseni]

Tableau (49) EKeGusii realization of the input erokeseni

This tableau shows that the optimal candidate is (b) even though it violates two relatively

low ranked constraints in EkeGusii grammar that is by adding of new segments and

changing of the features of segments in the loanword These violations however are not as

serious as the violation of maintaining complex vowels in the loan Therefore EkeGusii

constraint ranking prevails upon that of English English output of the same word will be

analyzed as in tableau (10)Input lǝʊkeɪintn locationOutput [lǝʊkeɪintn]Constraint ranking IDENT IO(F) MAX OI gtgt COMPLEX (V)

Input lǝʊkeɪintn IDENT IO(F) MAX OI (F) COMPLEX (V)

a [lǝʊkeɪintn]

b [erokeseni]

Tableau (410) English realization of the input lǝʊkeɪintn

Realization of the English diphthong ǝʊ as EkeGusii [o] in not peculiar to EkeGusii loans

from English only Other languages treat the diphthong the same way For example the

diphthong is nativized as [o] in Urdu and Punjabi loans from English as illustrated by (70)70) Urdu and Punjabi nativization of the English diphthong ǝʊa) Urdu realizationEnglish noun pronunciation Urdu realization pronunciationHotel hǝʊtǝl hootel [hotǝl]Coach kǝʊtint kooch [kotint]

b) Punjabi realizationEnglish noun pronunciation Urdu realization pronunciationRoad rǝʊd rood [rod]Coach kǝʊtint kooch [kotint]

(70) shows that in both Urdu and Punjabi the English diphthong ǝʊ is substituted for [o]

just like in EkeGusii Hussain (2011) attributes this realization to the fact that Urdu and

Punjabi phonological systems lack the diphthong meaning that it will be substituted for

that which is in the target language that is closest in terms of features This is indeed the

situation in this study EkeGusii phonological inventory lacks the diphthong [ǝʊ] which

leads for its substitution for [oo] which as has already been observed is phonetically closer

to the diphthong Other languages with similar results include Dholuo Owino (2003)

KiKamba Mutua (2007) Tonga Zivenge (2009) and Kalanga Chebanne amp Phili (2015)

among others These studies differ with the present one in two crucial ways some employ

different theoretical approaches (Owino 2003 amp Zivenge 2009) Chebanne amp Phili (2015)

like Hussain (2011) does not employ any theory while Mutua (2007) like the present

study employs Optimality Theory These studies focused on different languages

Nativization of English aɪ into EkeGusii [ae] ( aɪrarr[ae])

This is one of the cases where an English diphthong is substituted for an EkeGusii one The

diphthongs are characterized by the same initial element that is [a] which is [+LOW] The

second elements [ɪ] for English and [e] for EkeGusii though different share most feature

values The features are [-LOW -ROUND -BACK] This is perhaps the reason why the

second elements are substitutable [ɪ] becomes [e] because for one it is not present in

EkeGusii phonology Secondly it is the closest vowel to [e] in terms of phonetic features as

has already been observed (71] gives cases of realizations of English aɪ as EkeGusii [ae]

71) Nativization of English aɪ to EkeGusii [ae]English noun pronunciation EkeGusii nativized form pronunciationfile faɪl ebaeri [eaeri]tile taɪl etaeri [etaeri] mile maɪl emaeri [emaeri]

styile stail esitaeri [esitaeri]In (71) the English diphthong aɪ is realized as [ae] in EkeGusii nativized forms In

producing the diphthong aɪ in English the gliding begins with an open vowel which is

low central [a] and moves upwards to the high front position of [ɪ] It is one vowel which

in the given monosyllabic words serves as the peak of the syllable Its nativized form

however does not follow this form of gliding Instead the tip of the tongue is lowered to

the front mid close position of [e] as illustrated by charts (7) and (8)

a Chart (7) Production of the English diphthong [aɪ](Adapted from Roach 198320)

Chart (8) Production of EkeGusii nativized form [ae] from the English diphthong [aɪ]Adapted from Roach (1983 20)

The forms realized in (71) are phonetically distant from the possible source form [ai] not

only in in terms of gliding but also syllabically While in the source language the

combination is a diphthong in target language the resulting combination constitute of two

separate monophthongs each in its own syllable Thus the diphthong is monophthongized

For example [etaeri] from English taɪl lsquotilersquo there is no diphthong in EkeGusii

realization [etaeri] instead the English diphthong is split into two partseach part

forming a syllable of its own In fact the second part of the English diphthong[ɪ] is

changed to [e] which forms a sinle syllable in the nativized form

Other languages such as Dholuo (Owino 2003) and Kalanga (Chebanne and Phili 2015)

also nativize the English diphthong [aɪ] to [ae] (72) shows how Dholuo nativizes the

English [aɪ]

72 Nativization of English aɪ into Dholuo [ae] English noun pronunciation Dholuo nativized form pronunciationFile faɪl fael faelStyle staɪl stael staelMile maɪl mael maeltie taɪl tael taelSource Owino (2003)In Dholuo unlike in EkeGusii and indeed other African languages as (72) shows the

diphthong is treated as a single unit like in English and not as distinct vowels in separate

syllables This as will be discussed under phonotactic nativization is because EkeGusii a

Bantu language unlike Dholuo a Nilotic language strictly does not allow codas

Nativization of English eɪ into EkeGusii [e] (eɪ rarr [e] [ɛ])

The first element of the English diphthong and the EkeGusii vowel with which it is

substituted are similar phonetically The nativized form is created by dropping the final

element and lengthening the first element which is present in both phonologies as shown in

(73) Nativization of English eɪ to EkeGusii [e] or [ɛ]

English noun Pronunciation EkeGusii Nativized Form Pronunciationcake keɪk ekeeki [ekɛki]lsquocakersquocase keɪs ekeesi [ekesi]lsquocasersquobasin beɪsn ebeeseni [ɛɛseni]lsquobasinrsquostation steɪintn esiteseeni [esiteseni]lsquostationrsquo

In (73) the English diphthong eɪ is realized as [e] or [ɛ] in EkeGusii depending on the

vowels of the roots which harmonises with the rest of the vowels in the word This

involves the deletion of the second element of the diphthong and lengthening the first

element as in eɪ rarr [e] and changing the elements of the diphthong altogether and

adding length to the new element as in eɪ rarr [ɛ] These occurrences which are both

phonological and phonetic (LaCharite and Paradis 2003) are not confined to EkeGusii

phonology Languages for example Dholuo (Owino 2003) KiKamba Mutua (2007)

Punjabi and Urdu Haussin (2011) and Kalanga (Chabanne and Phili 2015) are

characterized by the same occurrences In KiKamba for example all cases of English eɪ

are realized as [e] as in (74)74) KiKamba realization of the English diphthong eɪEnglish noun pronunciation KiKamba nativized form pronunciationframe freɪm bulemu [ulemu]crane kreɪn keleni [keleni]crate kreɪt keleti [keleti]Adapted from Mutua (2007)

42113 Nativization of English triphthongs

A triphthong is defined as a vowel made up of three short vowels produced as one

(OrsquoConnor 1967 Roach 1983) The three vowels are treated as one because they are

treated as a single unit forming the syllable nuclei in the language (Chabanne and Phili

2015) This subsection shows how these triphthongs are realized in EkeGusii

As observed in section 4111 there are five triphthongs in English eɪǝ aɪǝ ͻɪǝ aʊǝ ǝʊǝ

Not many loaned words with these triphthongs were realized by the speakers While most

of the diphthongs were realized in very few cases others were not realized at all (75)

shows how these triphthongs were realized75)Nativization of the English triphthongs i) Realization of ǝɪǝ as EkeGusii [aja]

English noun pronunciation EkeGusii nativized form pronunciation

wire wǝɪǝ egwaya [e-γwaja]wire fǝɪǝ efaya [efaja] lsquowhoir kwǝɪǝ ekwaya [ekwaja]ii) Realization of eɪǝ as EkeGusii [aja]English noun pronunciation EkeGusii nativized form pronunciationplayer pleɪǝ epureya [epureja]layer leɪǝ ereya [ereja]iii) Realization of ͻɪǝ as EkeGusii [aja]English noun pronunciation EkeGusii nativized form pronunciationemployer emplͻɪǝ eemburoya [eemburoja]iv) Realization of aʊǝ and ǝʊǝ as EkeGusii [awa]English noun pronunciation EkeGusii nativized form pronunciationflower flaʊǝ eburawa [eurawa]shower intǝʊǝ esawa [esawa](75) shows that English triphthongs with the middle vowel being the front close vowel ɪ

are substituted for the form [aja] in EkeGusii while those which have the middle vowel

being the back high close vowel ʊ are substituted for EkeGusii [awa] Thus the English

triphthongal realization is lost in both cases Instead an approximant is introduced to

replace the middle element of the triphthong thus creating an extra syllable The

introduced approximant is determined by the backness and or the roundness of the vowel

The round vowel ʊ is replaced by the labial approximant [w] while the non-rounded

vowel ɪ is replaced by the palatal approximant [j] Both approximants unlike the vowels

they replace are present in EkeGusii phonology They are respectively closely related

phonetically to the vowels they replace This explains why the approximants are chosen

during nativization Figure (28) for the English word wǝɪǝ lsquowirersquo for example illustrates

how triphthongs are handled by EkeGusii OslashOslash w ǝ ɪ ǝ English

OslashOslash OslashOslashOslashOslash EkeGusii OslashOslashw a j a Phonemic substitution

e γ w a j a (Nativization Borrowing complete)Figure (28) Substitution of the English ǝɪǝ for EkeGusii [aja]

This figure shows that the English triphthong ǝɪǝ is realized as EkeGusii aja in which

the English short vowel ǝ is replaced with EkeGusii a and the English vowel ɪ is

replaced by the semi consonant j EkeGusii phonology like many other phonologies does

not have any triphthongal glide Thus any triphthong that comes into it is likely to take a

different form as it does in (75) Again the phonotactics of EkeGusii does not allow any

form of vowel clusters The vowel clusters of three as in the words in (91) have to be

declusterized to the acceptable phonotactic form as will be discussed under phonotactics

Nativization of the English triphthong by vowel declusterization through a replacement of

the medial vowel of the triphthong by an approximant is a common phenomenon For

example in Dholuo (Owino 2003) nativization of the English triphthong ǝɪǝ behaves

exactly the same way as in EkeGusii even though the two languages are from totally

different families (EkeGusii is Bantu while Dholuo is Nilotic) as illustrated by (76)76) Dholuo nativization of the English triphthong wǝɪǝEnglish noun Pronunciation Urdu Realization Pronunciationwire wǝɪǝ gwaya [waja]choir kwǝɪǝ kwaya [kwaja]

(76) shows that the triphthong is done away with by introducing a glide which takes the

position of the medial vowel This further leads to resyllabification of the word changing

from being monosyllabic to disyllabic This is indeed what happens to the diphthong in

EkeGusii nativization

However not all languages nativize the English triphthong by declusterization through

approximant introduction Urdu and Punjabi (Hussain 2011) for example nativize the

English triphthong ǝɪǝ by substituting it for [ae] a diphthong as shown in (77)

77) Substitution of English aɪǝ for Urdu and Punjabi [ae] i)UrduEnglish noun pronunciation Urdu realization Pronunciationfire faɪǝ faer [faer]wires waɪǝs waeles [waeles]

ii) PunjabiEnglish noun pronunciation Punjabi realization pronunciationfire faɪǝ faer [faer]diary daɪǝrɪ daer [daer]Adapted from Hussain (2011)

In (77) Urdu and Punjabi nativize the English triphthong aɪǝ as [ae] The first element in

the triphthong is maintained the medial element deleted while the final element is changed

from a mid-front vowel to a low front vowel These changes are different from those

observed in EkeGusii and Dholuo nativization of the triphthong In these cases the initial

and the final elements of the triphthong are maintained while the medial element is

replaced with an approximant All these occur as Owino (2003) observes in order to create

an acceptable syllable structure (to be discussed in detail under nativization by

resyllabification) In the case of EkeGusii and Dholuo for example the introduction of an

approximant to replace a vowel is intended to break the complex vowel that is not

acceptable in grammar of the languages and in the process an extra syllable is created

Punjabi and Urdu seem to tolerate a complex vowel a diphthong but not a triphthong It

deletes the last element which seems to be replaced with the alveolar tap [r] which closes

the syllable Thus Punjabi and Urdu tolerate syllable codas

In Optimality Theory perspective the realizations witnessed in (75)ndash (77) are accounted for

by the changes that take place As it has already been observed above English triphthongs

are realized differently in EkeGusii loaned words like in other language loaned words

There is loss of the English triphthongal status due to the replacement of the middle vowel

by an approximant which creates an extra syllable This change presupposes the following

OT constraints IDENT IO (F) which demands that features of an input segment must be

preserved in the output no feature change COMPLEX V which demands that complex

vowels are not allowed and MAX IO which demands that output segments must have input

correspondents no segment addition These constraints are ranked differently depending on

the output required EkeGusii does not allow complex vowels or clusters therefore it ranks

the markedness constraint higher than the faithfulness constraints Thus its ranking is

COMPLEX V gtgt MAX IO IDENT IO (F) English on the other hand tolerates

complex or vowel clusters meaning that the markedness constraint is dominated by the

faithfulness constraints Thus the constraints are reranked as follows IDENT IO (F)

MAX IO gtgt COMPLEX Given these rankings analyses of EkeGusii and English

realizations of the English triphthong aɪǝ for example are given in tableaux (11) and (12)

respectivelyEkeGusii realizationEkeGusii input waja

Input waja COMPLEX V MAX OI IDENT IO (F)

a [waɪǝ]

b [waja]

Tableau (411) Ekegusii realization of the input waja In this tableau candidate (a) loses to candidate (b) because (a) disobeys the higher ranked

constraint by allowing a complex vowel which is banned in this language candidate (b) on

the other hand wins because it obeys the determining constraintEnglish realizationInput waɪǝ lsquowirersquoOutput waɪǝ

Input waɪǝ IDENT IO (F) MAX OI COMPLEX V

a [waɪǝ]

b [waja]

Tableau (412) English realization of the input waɪǝ

In tableau (12) candidate (a) is the output because it obeys the highest ranked constraint as

compared to candidate (b) which violates the determining constraint

In essence nativization of English diphthongs and tripthongs in EkeGusii loaned nouns

from English generally involve monophthongization Monophthongization does not only

affect EkeGusii loans from English but other languages too For example languages such

as Kalanga spoken in Botswana (Chebanne amp Phili2015) KiKamba (Mutua 2013)

Tonga spoken in Zimbabwe (Zivenge 2009) and Dholuo spoken in Kenya (Owino

2003) among others behave the same way that is they get monophthongized For

example in KiKamba like EkeGusii Kalanga and Tonga (all Bantu) there are no

diphthongs or triphthongs in the strict sense of English KiKamba therefore like these

other Bantu languages monophthongizes any diphthong and triphthong that enters into its

phonology from English as in (78)

78) Monopthongization of diphthongs and triphthongs in KiKamba

Diphthong English Word Pronunciation KiKamba Realization Pronunciation

eɪ frame freɪm vulemu [ulemu]aɪ bicycle baɪskl vasikili [asikili]iǝʊ radio rediǝʊ letiu [letio]aʊǝ towel taʊǝwǝl taulo [taulo]Adapted from Mutua (2013)

(78) shows that English diphthongs and triphthongs are realized as monophthongs in

KiKamba This is how they are treated in EkeGusii as discussed above This is expected

because KiKamba like EkeGusii is a Bantu language and therefore share common

phonological features However analysis of Dholuo a non Bantu language show that

English diphthongs and triphthongs are equally monophthongized (Owino 2003) (79)

demonstrates this observation

79) Dholuo monophthongization and triphthongization

Diphthongtriphthong English noun pronunciation Dholuo realization pronunciation

ǝʊ coat k ǝʊt koti [koti]eɪ grade greɪd giredi [giredi]aʊ scout skaʊt sikaot [sikaot]aɪ file faɪl fael [fael]ɪa gear gɪa giya [gija]ǝɪǝ wire wǝɪǝ waya [waja]Adapted from Owino (2003) Monophthongization of English diphthongs and triphthongs in (79) mean that

monophthongization is not confined to Bantu languages only but rather that all those

languages without them irrespective of their language families

4212 Nativization of English consonants

Nativization consonants in English nouns borrowed from EkeGusii is done by replacing or

substituting the English consonant segments not present in EkeGusii phonological system

A number of consonant segments found in the phonological system of English do not exist

in EkeGusii phonology These include f v l Ө eth ʒ int h As it has already been

observed in this study the voiced plosives d g and b only occur with nasals

homorganistically in EkeGusii and are effectively regarded as pre-nasals It has also been

observed that the plosive p is only found in one or two idiophones according to Whiteley

(1960) Cammenga (2002) suggests that this plosive is ldquoincreasingly noticeable in the

speech of the younger generation which has had contact with Swahili and Englishrdquo This

study argues in favour of the fact that p is a rare sound in EkeGusii and that if all the

younger generation of the 1960s when Whitely conducted his research on the language

were using it it was only in nativized words from the languages mentioned by Whiteley

This section shows how the English consonants (listed above) not found in EkeGusii

phonological structure are realized by EkeGusii speakers The realizations are accounted

for within Optimality Theory perspectives As it has already been mentioned nativization of English consonants in EkeGusii involves

change or substitution of English consonants for those of EkeGusii This in Optimality

Theory means that there is a violation of a faithfulness constraint - IDENT IO (SEG) C or

IDENT IO(F) or IDENT IO (P) and IDENT IO (VOICE) which demands that an input

consonant segment or feature specification of a segment or place of articulation of a

segment and voice of a segment must have an output correspondent respectively This

demand however would satisfy a number of markedness constraints such as VOI which

prohibits voiced obstruents like [g d z] VTV which bans voiceless obstruents in

intervocalic positions as in [ota] versus [oda] McCarthy (2007) AGREE (VOICE)

AGRREE (CONTINUANT) SRIDENT (FRIC) which prohibits noisy fricatives

(Laparombara 2013) The markedness constraint therefore would be ranked higher than the

faithfulness constraint in these realizations because faithfulness constraints will be violated

at their expense Thus markedness constraints will dominate faithfulness ones

MARKEDNESS gtgt FAITHFULNESS This is the ranking which prevails in the

realizations of English consonants loaned into EkeGusii Re-ranking of the constraints so

that the faithfulness constraints dominate the markedness ones give opposite results That

is English outputs are realized

42121 Nativization of English f and v

The English labiodental fricatives share the labial feature value with EkeGusii [β] which

they are substituted for The feature value that separate the English consonants from that of

EkeGusii are while f v are [+DENTAL] and [+ STRIDENT] [β] is [ndashDENTAL] and [-

STRIDENT] It is the labial feature which they share which probably make them

substitutable as shown in (80)

80) Realization of English f andv by EkeGusii [β] [ɸ] English noun pronunciation EkeGusii nativized form pronunciationverandah vǝraeligndǝ ebaranda [eβaranda]tv tivi etibii [etiβii]fashion faeligintn ebaasoni [eβaasoni]fridge frɪdʒ eburichi [euritinti]

(80) shows that the English voiced labiodental fricatives v and its voiceless counterpart f

are realized as [β] a voiced bilabial fricative in EkeGusii The possible explanation to this

is that EkeGusii phonology lacks these anterior strident labiodental fricatives and therefore

the anterior non strident bilabial fricative substitutes them because of phonetic similarity

between them They share [+LABIALITY] and [+CONTINUANCY] differing majorly in

terms of stridency Therefore the realization of v and f as [β] is as a result of phonetic

similarity Other languages nativize the English v and f in a similar manner In Dholuo

for example v becomes [b] (Owino 2003) This is even when the phonetic distance

between the two obstruents seem to be more than it is in EkeGusii The consonant

substitution process that takes place in data set (80) is illustrated by figure (29) which

considers nativization of the English noun fridge frɪdʒ to EkeGusii eburichi [euritinti]

Oslashf Oslash r ɪ dʒOslash English OslashOslashOslash r Oslash OslashOslash EkeGusii

Oslash Oslash r i tintOslash Phonemic substitution

e u r i tint i (Nativization Borrowing complete)Figure (29) Substitution of the English f for EkeGusii [β]

This figure shows that the English consonant f is substituted for the EkeGusii consonant

In OT this realization would be accounted for by the fact the English segment f a

labiodental strident fricative changes to []a bilabial non-strident fricative in EkeGusii

segment This suggests that a segment in the output will not be faithful to its input form

presupposing the faithfulness constraint IDENT IO (SEG) C It also presupposes that

strident fricatives are not allowed in outputs yielding the markedness constraint

[STRIDENT] (FRIC) Thus the markedness constraint will therefore be ranked higher

than the faithfulness one in EkeGusii outputs because faithfulness is inconsequential in

determining the optimal candidate in the language The opposite ranking of the constraints

as shown in tableaux (13) and (14) yields English outputs EkeGusii realization of English f

EkeGusii input [eβuritinti] lsquofridgersquo

Constraint ranking [STRIDENT] (FRIC) gtgt IDENT IO (SEG) C

Input eβuritinti [STRIDENT] (FRIC) IDENT IO (SEG) C

a) frɪdʒ

b) eβuritinti

Tableau (413) EkeGusii realization of the input eβuritinti

Candidate (b) is the output because it violates a less consequential candidate in EkeGusii in

a bid to satisfy the demands of the highly ranked constraint [STRIDENT] (FRIC) which

prohibits strident fricatives in EkeGusii Candidate (a) violates this highly ranked constraint

in the language the reason why it loses This is compared to the English realization of the

same word as follows

English input frɪdʒ lsquofridgersquo

Constraint ranking IDENT IO (SEG) C gtgt [STRIDENT] (FRIC)

Input frɪdʒ IDENT IO (SEG) C [STRIDENT] (FRIC)

a) frɪdʒ

b) eβuritinti

Tableau (414) English realization of the input frɪdʒ In this tableau candidate (a) is the output because it obeys the faithfulness constraint

IDENT IO (SEG) C which is ranked higher than the markedness constraint in English An

alternative explanation to the realizations given in (85) above can be given In Tonga

(Zivenge 2009) the English voiceless labiodental fricative f is realized as [v] a strident

labiodental just like f This is irrespective of the fact that the voiceless labiodental is

present in both the phonology of English and Tonga Zivenge attributes this occurrence to

the fact that the sound is constrained in terms of occurrence to onsets of the last syllable of

class 7 nouns in Tonga This occurrence however can be given another interpretation In

EkeGusii in which as data set (96) shows this labiodental fricative is realized as the voiced

bilabial fricative [] while in KiKamba loans from English Mutua (2013) it is realized as

[] a voiceless bilabial fricative In White Hmong loans from English Golston and Yang

(2001) the fricative is maintained irrespective of the position it occupies in a word

While it is agreeable that the labiodental takes different realizations in loanwords from

English depending on the phonology of the host languages it can be argued that whichever

segment they substitute in the target language is normally voiced intervocalically and

voiceless in voiceless environments

In Tonga (Zivenge 2009) the English word scarf ska f is realized as sikava [sikava] In

EkeGusii as (96) shows the word fridge frɪdʒ is realized as eburichi [euritinti] These

show that the two languages realize the sound f differently intervocalically [v] in Tonga

and [] in EkeGusii Besides these sounds sharing labiality and frication they are voiced

Given their environment of occurrence in the target languages ( are intervocalic) and their

voiced nature the VTV markedness constraint in Optimality Theory (McCarthy 2003) is

presupposed This constraint prohibits voiceless consonants in intervocalic positions The

realizations would also assume a number of faithfulness constraints such as IDENT IO

(SEG) C which demands that input consonant segments must have output correspondents

Thus the markedness constraint dominates the faithfulness one in the realization of The

English f in the above examples The following tableau for the English input scarf ska f

in Tonga realization illustrates this observationTonga input [sikava]Constraint ranking in Tonga VTV gtgt IDENT IO (SEG) C

Input skava VTV IDENT IO (SEG) C

a) skafu

b) sikava

Tableau (415) Tonga realization of the input skava

Candidate (a) in the tableau is optimal because it satisfies the highest ranked constraint

(VTV) (a) loses because it violates the constraintReranking the given constraints will yield English output of the given word since the

languages rank constraints differently

42122 Nativization of English eth and

The consonants [Ө] and [eth] are characterized by the feature values [+interdental] [+

continuant] [+coronal] and [+ anterior] They are distinguished by the feature [voice]

While [Ө] is voiceless [eth] is voiced The consonant [t] with which the given English

consonants substitute share a number of feature values such as [+anterior] and

[+coronal] which determine its choice for substitution EkeGusii does not have interdental

fricatives This is perhaps the reason behind the avoidance of the interdentals which are

marked in EkeGusii as (81) shows

81) Realization of the English Ө and eth as EkeGusii [t] English noun pronunciation EkeGusii nativized form pronunciationthermos Өɜməs etamosi [tamosi]thief θif etiβu [ tiβu]father f ɜethə omoβata [omoβata]The substitution process involved in these realizations is illustrated by figure (30) for the

English noun lsquothermosrsquo Өɜməs OslashӨ ɜ m ə s Oslash English OslashOslash Oslash m Oslash s Oslash EkeGusii Oslasht a m o s Oslash Phonemic substitution

e t a m o s i (Nativization Borrowing complete)Figure (30) Substitution of the English Ө for EkeGusii [t]Adapted from Gussenhoven and Jacobs (2011)

This figure shows that the English Ө is substituted for Ekegusii [t] The voiced equivalent

of Ө that is eth is also substituted for [t] This is explained as follows Firstly Ekegusii

does not have the two interdental fricatives and therefore the nearest consonant in terms of

feature values that is [t] is the likely choice in this case As has already been observed [t]

is characterized as an alveolar in EkeGusii (Whitely 1960 and Cammenga 2002)

However as Cammenga (2002 54) suggests ldquothough t is characterized as an alveolar it

may be rendered as a voiceless interdental obstruent possibly also as an alveolo-dental or

perhaps dental-alveolar obstruentrdquo This study takes the position that the sound is an

alveolo-dental obstruent and therefore excludes the possibility that it may be an interdental

or a dental alveolar obstruent Thus [t] therefore is the likely choice as the substitute of the

two English interdentals in English words loaned into EkeGusii

Secondly Ekegusii language lacks voiced consonants (except pre-nasal stops nasals and

the fricative [ndʒ]) (Cammenga 2002) This means that chances of the English voiced

inter-dental eth being an output in the nativized English loaned words in EkeGusii are

minimal if at all

Given that interdentals are not realized in EkeGusii and that they are not part of the

phonological system of the language a markedness constraint of place of articulation

which bans interdentals is proposed that is INTERD (FRIC) (McCarthy 2003) This

means that interdentals are not allowed in this language in other words they are marked A

survey of literature seem to support this proposal In Tonga Zivenge (2009) English Ө is

realized as [s] as in theory Өɪǝrɪ rarr [sijori] while eth is realized as [dʒ] as in leather

leethǝ rarr [ledʒa] in Japanese Kay (1996) English Ө is realized as [s] as in thrill Өrɪl rarr

[siriru] among others In all these realizations there is avoidance of the English

interdentals Thus the proposal that the given interdental fricatives and the interdental

position generally are marked is supported Given that interdentals are avoided resulting to

a change of segment as shown in (97) the faithfulness constraint IDENT IO (SEG) C is

violated Tableaux (16) and (17) below analyses the effect of the ranking and re-ranking of

the constraints in English and EkeGusii respectively for the English word Өɜməs

lsquothermosrsquo

English input Өɜməs lsquothermosrsquoConstraint ranking in English IDENT IO (SEG) C gtgt [INTERD] (FRIC)

Input Өɜməs IDENT IO (SEG) C [INTERD] (FRIC)

a) Өɜməs

b) tamosi

c) ethɜməs

Tableau (416) English realization of the input Өɜməs In tableau (416) (a) is the output because it obeys the faithfulness constraint which ranked

higher than the markedness constraint (b) and (c) on the other hand lose because they

violate the constraint This is comparable to EkeGusii ranking of the constraints as followsEkeGusii input tamosi lsquothermosrsquoConstraint ranking in EkeGusii [INTERD] (FRIC) gtgt IDENT IO (SEG) C

Input tamosi [INTERD] (FRIC) IDENT IO (SEG) C

a) Өamosi

b) tamosi

c) ethɜməs

Tableau (417) EkeGusii realization of the input tamosi

This tableau shows that ranking of the markedness constraint over the faithfulness one

yields EkeGusii output since EkeGusii ranks the faithfulness constraint lower than the

markedness constraint

42123 Nativization of English lThe English consonant l is realized as [r] in EkeGusii [l] and [r] are characterized by the

following feature values [+ANTERIOR] [CORONAL] among other features They are

distinguished by the feature lateral while [l] is [+LATERAL] [r] is [-LATERAL]

meaning that it is a trill EkeGusii phonology does not have the lateral consonant Given

that the two consonants are closely related in terms of feature values substituting [l] for [r]

is natural as in (82)82) Substitution of English l by EkeGusii [r] (l rarr [r]) English noun pronunciation EkeGusii nativized form pronunciationtelevision tεlεviintn eterebisoni [tereβisoni]class klaeligs ekerasi [ekerasi]

lunch lʌndʒ ranchi [rantinti]glucose glukǝʊz gurukosi [γurukosi]In (98) the English consonant l is substituted for [r] in EkeGusii The substitution process

involved is illustrated by the English word klaeligss lsquoclassrsquo in figure (31) Oslash k l aelig s Oslash English

Oslash k OslashOslash Oslash s Oslash EkeGusii Oslashk Oslash r a s Oslash Phonemic substitution

e k e r a s i (Nativization Borrowing complete)Figure (31) Substitution of the English l for EkeGusii [r]The consonant l is not found in EkeGusii phonological system This explains why it is not

realized in data (82) Substitution of [r] for the English l by second language speakers is a

common phenomenon For example in Japanese Kay (1996) like in EkeGusii the English

l is realized as [r] as in (83)83) substitution of English l for Japanese [r]English noun pronunciation Japanese nativized form pronunciationhustle hʌsl hassuru [hasuru]whistle hwɪsl hoissuru [hoisuru]last laeligst rasuto [rasuto]slip slɪp surippu [suripu]Adapted from Kay (1996)In Japanese like in EkeGusii the alveolar lateral [l] is not found in its phonological system

(Kay 1996) That is why it is substituted for the alveolar trill as in (82 amp 83) above which

is closer to it in terms of phonetic features as has already been observed

Some languages however behave in the exact opposite of what happens in EkeGusii and

Japanese regarding these two segments In Hawaiian (Golston amp Yang 2001) KiKamba

(Mutua 2007) and Tonga (Zivenda 2009) among others for example English r is

realized as [l] Thus in Hawaiian merɪ merry rarr [mele] in Tonga rulǝ ruler rarr [lula]

while in KiKamba krim cream rarr [kelimo] In these languages unlike EkeGusii and

Japanese [r] is not found in their phonological systems

What these substitutions of the English l for [r] in target or borrowing languages and the

other way round mean is that one of these approximants normally occur in a language and173

not both In other words it is normally only either of them which occurs in a given

phonological system and not both This however does not close out the possibility of some

languages having both of these approximants for example in Kalanga (Chebanne and

Phili 2015) just like in English both [l] and [r] are present in its phonological system

That is why the approximants are retained in Kalanga nativized forms from English as in

84) Kalanga nativization of the English approximants l and rEnglish noun pronunciation Kalanga nativized form pronunciation

driver draɪvǝ dirayivara [dirajivara]plastic plaeligstɪk pulasitiki [pulasitiki]plate pleɪt puleyiti [pulejiti]film fɪlm filimu

Adapted from Chebanne amp Phili (2015)

[filimu]

In this data the English sounds r and l are retained in Kalanga nativized forms of

English The explanation that can be given as to why these approximants are substitutable

cross linguistically is that they share all but only one phonetic feature they are produced at

the alveolar ridge they are approximants and voiced But while [l] is a lateral [r] is a trill

The cross linguistic alternation between these segments as observed above is a common

phenomenon (John 1984) Given this kind of alternation in which some languages prefer

the lateral consonant while others the trill in their output forms a manner of articulation

markedness constraint can be presupposed Such a constraint may read Languages that

have trills do not allow laterals (LATERAL) languages that have laterals do not allow

trills (TRILLS) (McCarthy 2003) The constraint LATERAL therefore bans trills while

TRILL disallows laterals in outputs These markedness constraints presuppose

faithfulness constraints such as IDENT IO (SEG) C

EkeGusii which lacks the lateral sound handles the English input ekerasi class as analyzed

in tableau (18)Constraint ranking LATERAL gtgt IDENT IO (SEG) C

Input ekerasi LATERAL IDENT IO (SEG) C

a) claeligs

b) ekerasi

c) ekelasi

Tableau (418) EkeGusii realization of the input ekerasi Candidate (b) is the output even when it violates the faithfulness constraint Itrsquos being

optimal is motivated by the constraint LATERAL which bans laterals in EkeGusii and

which is ranked higher than the faithfulness constraint (a) and (c) are not optimal because

they violate this constraint

Languages like KiKamba and Tonga among others that do not have the trill segment [r]

would have a different analysis as tableau (19) for KiKamba realization of the input skulu

lsquoschoolrsquo showsConstraint ranking in KiKamba TRILL gtgt IDENT IO (SEG) C

Input skulu TRILL IDENT IO (SEG) C

a) sukuru

b) sukulu

Tableau (419) KiKamba realization of the input skulu Candidate (b) is optimal in this tableau because it obeys the constraint that is relatively

highly ranked in KiKamba that is TRILL suggesting that KiKamba does not allow trills

42124 Nativization of English z

The common feature values that characterize these consonants include [+OBSTRUENT]

[+CONTINUANT] and [+ANTERIOR] among others They are distinguished by the

feature [VOICE] While [z] is [+VOICE] [s]is [-VOICE] EkeGusii does not have the

voiced obstruent which explains why it is not realized in the loaned words from English as

in (85)

85) Substitution of English z by EkeGusii [s] English noun pronunciation EkeGusii nativized form pronunciationzero zirəʊ esiro [siro]zone zͻn esoni [soni]maize meɪz ebimeisi [meisi]gas gᴂz egasi [eγasi]The substitution process involved in this realization is illustrated by the English word

zirəʊ lsquozerorsquo in figure (32) Oslash z i r ǝʊ English OslashOslash i r Oslash EkeGusii Oslashs Oslash r o Phonemic substitution

e s i r o (Nativization Borrowing complete)Figure (32) Substitution of the English z for EkeGusii [s]

Substitution of the English z for [s] in loanwords affects other languages too especially

those that lack in their phonological systems In Kikamba (Mutua 2007)for example

fees fiz rarr viisi [iisi] A similar occurrence is witnessed in Dholuo (Owino 2003) as in

lsquogazettersquo gǝszet rarr [gaset] This is even when the two assimilating languages are from

different language families KiKamba being Bantu while Dholuo being Nilotic

The substitution of [s] for [z] can be explained by the fact that the segments share all but

one feature They are [+OBSTRUENT] [+CORONAL and [+ANTERIOR] among other

features They are distinguished by the feature voice While [s] is voiceless [z] is voiced

EkeGusii realization of [s] for English z presupposes the OT markedness constraint

VOICEOBS (McCarthy 2003) which prohibits voiced obstruents The segmental change

assumes the faithfulness constraint IDENT IO (SEG) which is violated at the expense of

the markedness constraint which must not be violated Therefore VOICEOBS dominates

IDENT IO (SEG) in EkeGusii and the other way round in English This is analyzed in

tableaux (20) and (21) for the English input esiro and zirəʊ lsquozerorsquo for EkeGusii and

English respectively Constraint ranking in EkeGusii VOICEOBS gtgt IDENT IO (SEG)

Input esiro VOICEOBS IDENT IO (SEG) C

a) zirəʊ

b) esiro

Tableau (420) EkeGusii realization of the input esiro Candidate (b) is the output because it obeys VOICEOBS which is highly ranked in

EkeGusii (a) on the other hand violates the constraint and therefore loses English reranks

the given constraints differently from EkeGusii ranking with different results The

reranking of the constraints will give the analysis of tableau (21)Constraint ranking in English IDENT IO (SEG) gtgt VOICEOBS

Input zirəʊ IDENT IO (SEG) VOICEOBS

a) esiro

b) zirəʊ

Tableau (421) English realization of the input zirəʊ This tableau shows that (b) is the output English ranks the faithfulness constraint higher

than the markedness constraint Thus the demand of having corresponding segments

between inputs and outputs in English is more serious than allowing different outputs

English allows many voiced obstruents in its outputs in order to be faithful to the inputs

42125 Nativization of English int and ʒ

These consonants share the feature values [+CORONAL] and [+CONTINUANT] among

others But while int and ʒ are [-ANTERIOR] the consonant [s] is [+ANTERIOR] Thus

EkeGusii does not allow non-anterior coronal continuants explaining why they are avoided

by speakers as in (86)

86) Realization of English int as EkeGusii [s] English noun pronunciation EkeGusii nativized form pronunciationshow intǝʊ esoo [esoo]shirt intɜt esati [esati]shock intͻk esoki [esͻki]bishop bɪintͻp ebisobu [eisͻu]The substitution process involved in the realizations in these data set is illustrated by figure

(33) for the English word intɜt lsquoesatirsquo Oslash int ɜt Oslash English

OslashOslash Oslash r Oslash EkeGusii Oslashs a t Oslash Phonemic substitution

e s a t i (Nativization Borrowing complete)Figure (33) Substitution of the English int for EkeGusii [s]

(186) indicates that the English consonantal segment int is not realized by EkeGusii

speakers It is not found in EkeGusii phonological system It is substituted for by the

segment [s] which is in terms of phonetic features closest to it They share the features [+

CORONAL + CONTINUANT +CORONAL] However while [s] is [+ ANTERIOR] [int]

is [-ANTERIOR] This points to the fact that the feature [anterior] determines the choice of

[s] and not [int] in EkeGusii realization Thus EkeGusii does not allow [-anterior +coronal

+continuant] in its outputs These features combine to form a markedness constraint of the

form NON-ANTERCORLFRIC (McCarthy 2003)) which prohibits non anterior coronal

fricatives in EkeGusii Like other markedness constraints this constraint presupposes the

faithfulness constraint IDENT IO (SEG) because it involves a change of an input segment

in its output form This is analyzed in tableaux (22) and (23) for the inputs esati and intɜt

lsquoshirtrsquo for EkeGusii and English respectivelyEkeGusii realization

Constraint ranking NON-ANTERCORLFRIC gtgt IDENT IO (SEG)

Input esati NON-ANTERCORLFRIC IDENT IO (SEG)

a) intɜt

b) esati

Tableau (422) EkeGusii realization of the input esati English RealizationConstraint ranking IDENT IO (SEG) gtgt NON-ANTERCORLFRIC

Input intɜt IDENT IO (SEG) NON-ANTERCORLFRIC

a) intɜt

b) esati

Tableau (423) English output and input of intɜtIn tableau (22) the output is candidate (b) because it obeys the relatively high ranked

constraint in EkeGusii The reranking of the constraints in English yields a different output

in tableau (23) which obeys the faithfulness constraint These analyses point to the fact that non anterior coronal continuants are marked This

observation is supported by similar realization of the English int loaned words in other

languages such as in Dholuo as in (87)

87) Nativization of English int in Dholuo English noun pronunciation EkeGusii nativized form pronunciationshirt intɜt sati [sati]fashion faeligintn fason [fason]bishop bɪintͻp pisopu [pisopu]

Source Owino (2003)In (87) the English consonant int is realized as [s] in Dholuo Thus like in EkeGusii the

consonant is equally marked in Dholuo

42126 Nativization of English g in EkeGusii

The consonants [g] and [γ] are closely related in terms of feature values They are [-

CORONAL] [+ VELAR AND [-ANTERIOR] They are distinguished by the feature

[CONTINUANT] While [γ] is [+CONTINUANT] [g] is [-continuant] This explains why

EkeGusii realizes the English g as [γ] as in (88)88) Nativization of English g as EkeGusii [γ]English noun pronunciation EkeGusii nativized form pronunciationbag bᴂg epagi [epaγi] glue glu eguru [eγuruu] goal gǝʊl egori [eγorigroup grup egurubu [eγuruu] In (88) the English velar non-continuant g is substituted for the voiced EkeGusii Velar

fricative [γ] in order to be dissimilar in terms of features with the consonant in the adjacent

syllable that is [p] While [γ] is [+ CONTINUANT] and [+VOICE] [p] is [-

CONTINUANT] and [-VOICE] Thus Dahlrsquos Law of feature dissimilation described

elsewhere in this study which affects EkeGusii is in operation here Figure (34) for the

English word bᴂg lsquobagrsquo is an illustration of this observation Oslash baelig g Oslash English OslashOslash OslashOslashOslash EkeGusii Oslashp a γ Oslash Phonemic substitution

e p a γ i (Nativization Borrowing complete)Figure (34) Substitution of the English g for EkeGusii [γ]

In figure (34) the English voiced velar non continuant g is substituted for by EkeGusii

[γ] a voiced velar continuant In essence the stop is fricativized This in Optimality

Theory presupposes that in EkeGusii assimilation of certain phonemic features such as

continuancy and voice are not allowed in a word This is feature dissimilation which in OT

can be put as a markedness constraint that disallows assimilation of features in a domain

such as a syllable or a word that is ASSIM (F) (McCarthy 2002) This constraint

presupposes that phonemes in the output forms of a word will not be the same in their input

correspondents the result being a faithfulness constraint such as FAITH (F) Using these

constraints the realization of the English word baeligg lsquobagrsquo is analyzed in tableaux (24) and

(24) for EkeGusii and English rankings respectively EkeGusii input epeγi lsquobagrsquoconstraint ranking [ASSIM] (F) gtgt FAITH (F)

Input epaγi [ASSIM] (F) FAITH (F)

a) baeligg

b) epaγi

Tableau (424) EkeGusii realization of the input epaγi

In this tableau candidate (b) is optimal because it does not assimilate features across the

syllables of the word In candidate (a) which loses there is assimilation of the features

voice and stop in the word This is not allowed in EkeGusiiEnglish input baeligg lsquobagrsquoConstraint Ranking FAITH (F) gtgt [ASSIM] (F)

Input baeligg FAITH (F) [ASSIM] (F)

a) baeligg

b) epaγi

Tableau (425) English realization of the input of baeliggCandidate (a) is the output because English ranks the faithfulness constraint higher than the

markedness one which disallows feature assimilation

Besides the kind of substitution of the consonants described so far there is also substitution

of the voiceless English pre-nasal consonants k and s with their voiced equivalents in

EkeGusii that is gand z respectively and substitution of the voiceless English stop k

with EkeGusii voiced fricative [γ] This is Dahlrsquos law in operation (Bickmore 1998) This

process is discussed in detail under phonological processes in sub-section 4241

422 Phonotactic nativization

In section 421 above segmental nativization of English nouns in EkeGusii which is one

of the areas of focus in phonological nativization has been analzed and discussed In this

section the second area Phonotactic nativization is analyzed and discussed Phonotactic

nativization in this study focuses on the syllable structures of the two languages in contact

Kager (199991) observes that the syllable is a useful unit of phonological analysis it is in

the syllable that the rules that govern the well formedness sequences of segments in a

language consonants and vowels are defined

Kagar (1999) observation that the syllable controls processes of insertion and deletion of

segments where either a cluster is not required or a consonantvowel is not required word

initially or finally and that the syllable links segments to higher- level (suprasegmental)

units forms the basis of analyses in this section The study shows syllable differences

between EkeGusii and English and how the English syllable adjusts syllabically to conform

to the phonotactic patterns of EkeGusii the target language In particular the study shows

how epenthesis breaks consonant clusters on syllable margins (onsets and codas) making

the syllables open in cases where the incoming words have a closed syllable It also shows

how deletion helps to achieve similar results (that is avoidance of closed syllables)

As was observed in subsection 413 EkeGusii is a (V) CV syllable structure while English

on the other hand is a (C) C) (C) V (C) (C) (C) (C) language This is illustrated by (89)

89) EkeGusii and English syllable structures

a) EkeGusii syllable structure

syllable structure noun pronunciation gloss

CV mama mama mother V CV omote omote tree

b) English syllable structure

Syllable structure word pronunciation

V owe ǝʊ CVC take teɪk CCVCC stains steɪns CCCVC structure strʌktintǝ CVCCC facts fᴂkts CCVCCCC glimpsed glɪmpstThese syllable types can be presented on syllable nodes as figures (35) and (36)

demonstratei) CV as in mama lsquomotherrsquo

O N O N

C V C V

m a m a

ii) VCV as in omote lsquotreersquoσ σ σ

N O N O N

V C V C V

o m o t e

Figures (35) EkeGusii syllable types

i) V as in ǝʊ lsquoowersquo

ǝʊ ii) CCVCC as in steɪns lsquostainsrsquo

CC V CC

s t eɪ nz

Figures (36) English syllable types

The figures in (35) show that EkeGusii syllables do not allow complex margins All

margins that is onsets and codas are simple English syllables as shown by figures in (36)

on the other hand allow complex margins both in the onset and coda positions

The syllable types in (89) as illustrated in figures (35) and (36) presuppose a number of OT

markedness constraints ONSET which demands that syllables must have onsets CODA

which disallows closed syllables COMPLEX C which demands that syllables must not

have complex margins or that syllables must not have a cluster of consonants (McCarthy

2003) They also presuppose the faithfulness constraints MAX IO which provides that

input segments must have output correspondents no deletion of a segment and DEP IO

(SEG) which provides that no segment should be added to the output that is no epenthesis

What differentiates the two syllable typologies in this study is the different ranking of the

constraints by the languages In (105a) for example ONSET is ranked higher than DEP IO184

(SEG) which is presented as ONSET gtgt DEP-IO Tableau (26) for the EkeGusii input

ma ma [mama] lsquomother lsquoillustrates

Input mama

Constraint ranking ONSET gtgt DEP-IO

Input mama ONSET DEP-IO

a) mama

b) a ama

c) maa

Tableau (426) EkeGusii realization of the input mama

In this tableau candidate (a) is optimal because it violates neither of the constraints

Outputs (b) and (c) lose because they violate the highly ranked constraint in the language

ONSET Thus a markedness constraint dominates a faithfulness one

The structures in (89b) on the hand allow the violation of ONSET in (i) NOCODA in (ii)

and COMPLEX in (iii) These constraints would be ranked as follows (i) MAX IO (SEG)

gtgt ONSET (ii) MAX IO (SEG) gtgt NO CODA (iii) MAX- I0 (SEG) gtgt COMPLEX

This is analyzed in tableaux (27-29)

i) Input ǝʊ lsquoowersquo

Constraint ranking in English DEP-I0 gtgt ONSET

Input ǝʊ MAX IO(SEG) ONSET

a) ǝʊ

b) tǝʊ

c) ǝʊǝ

Tableau (427) English realization of the input ǝʊ

ii) Input teɪk take

Constraint ranking MAX(seg) -IO gtgtNOCODA

Input teɪk MAX-IO(SEG) NO CODA

a) teɪ

b) teɪk

c) teɪki

Tableau (428) English realization of the input teɪk

iii) Input steɪnz lsquosteinsrsquo

constraint ranking MAX I0 (SEG) gtgt COMPLEX

Input steɪnz MAX-IO(seg) COMPLEX

a) seɪn

b) siteɪnsɪ

c) steɪns

Tableau (429) English realization of the input steɪnz

These tableaux show that faithfulness constraints are ranked higher than markedness

constraints Thus in English the demand that output forms be as similar as possible to their

input correspondents outranks the well formedness of outputs Therefore faithfulness

constraints dominate markedness constraints FAITHFULNESS CONSTRAINTS gtgt

MARKEDNESS CONSTRAINTS Thus violating a faithfulness constraint in English is

more serious than violating Markedness constraints EkeGusii on the other hand as was

realized in this study unlike English ranks markedness constraints higher than the

faithfulness constraints as illustrated by tableau (25)

Syllable nativization or syllabification is governed by syllable well formedness in the target

language (EkeGusii in this case) A loaned word in most cases violates some constraints of

syllable well-formedness in the new language This is because each language ranks

constraints (from the universal set) differently (Prince and Smolensky 1993) English loans

coming into EkeGusii have a different syllable structure and therefore have to undergo

syllable changes in order to be accommodated in the EkeGusii phonological system It was

discovered that EkeGusii language uses a number of strategies to change the syllable

structures of the English loans coming into its phonological system The speakers

employed vowel epenthesis (anaptyxis and paragogue) strategy to deal with consonant

syllable clusters and syllable codas This is analzed in section 4221 below

4221 Nativization by vowel epenthesis

Epenthesis according Katamba (1989) Eichoff (1990) Uffmann (2001 amp 2004) and Rose

amp Demuth (2006) among others is where new segments appear from zero in the formerly

unoccupied positions in a word or morpheme or between two previously abutting segments

There are three types of such insertions prothesis which is the insertion of an initial

segment normally a vowel anaptyxis which is an insertion of a vowel between two

consonants and paragogic insertion which inserts a segment at the end of a word

especially a vowel (Katamba 1989) This section considers paragogic and anaptyxistic

insertions which seemed to be the preferred forms of epenthesis in EkeGusii

Anaptyxis and paragogue were used by EkeGusii speakers to break complex syllable

margins and open closed syllables of the loans from English This is because EkeGusii

neither tolerates complex syllable margins nor codas or closed syllables Anaptyxis was

found to be common in handling of complex margins while paragogue was used to handle

codas in English nouns borrowed into EkeGusii The two processes are discussed in

sections 42211 and 42212 respectively

42211 Nativization of complex syllable margins

In section 4131 it was observed that English is characterized by complex syllable

margins of up to three consonants in onsets and four in codas (Cruttenden 2011) EkeGusii

on the other hand does not allow consonant clusters of any nature Its phonotactics is

characterized by simple syllable margins (Cammenga 2002) Thus any consonant clusters

in EkeGusii loans from English are broken through a number of ways such as anaptyxis as

shown in (90)

90) Nativization of English nouns borrowed into EkeGusii through anaptyxisEnglish noun pronunciation EkeGusii nativized form pronunciation store stͻ sitoo [sitoo]

brush brʌint burasi [βurasi]school skʊl sukuru [sukuru] glass glɑs kerasi [kerasi]

In (90) the English nouns contain a syllable with a complex margin while their EkeGusii

nativized forms do not The nativised forms were realized through among other processes

anaptytxis (vowel insertion) an epenthetic process This process is illustrated by figure

(37) for the word stͻ lsquostorersquo

English EkeGusii

σ σ σ

O N O N O N

CC V C V C V

s t ͻ s i t o

Figure (37) Nativization by anaptyxis

Figure (37) shows that the consonant cluster st in the English stͻ is broken by

insertion of the vowel [i] in EkeGusii to create the syllables [si] and another additional one

([to]) since the consonant t is pushed into an onset position of a new syllable Therefore

while there is only one syllable in the English word form there are two in the EkeGusii

nativized form

In Optimality Theory epenthesis leads to violation of a faithfulness constraint DEPndashI0 V

because the epenthetic segment has no counterpart in the input In the case of the loaned

word the well formedness embodied in COMPLX C among other constraints (ONSET

and NO CODA) dominate the DEP ndashI0 constraint (Uffiman 2004) Thus epenthesis is

opted for at the expenses of DEP- I0 V which is opposed by the syllable markedness

constraint In other words the cost of violating DEP ndashIO is less than that of the occurrence

of an impossible syllable structure in the native system

Following this argument the nativization process in (90) above for the word sku l

lsquoschoolrsquo for example is analyzed by tableaux (30) and (31) for English and EkeGusii

ranking of the constraints respectively

English input skul lsquoschoolrsquo

English constraint ranking DEP-I0 gtgt COMPLEX

Input skul DEP-IO COMPLEX

a) sukuru

c) skul Tableau (430) English realization of the input skul

EkeGusii input sukuru lsquoschoolrsquo

EkeGusii ranking COMPLEX gtgt DEP IO

Input sukuru COMPLEX DEP IO

a) skuru

c) sukuru

Tableau (431) EkeGusii realization of the input sukuru

In tableau (31) candidate (b) is optimal because it does not violate the comparatively high

ranked constraint in EkeGusii The candidate avoids complex syllable margins that are

banned in EkeGusii The reverse is true in English language as tableau (29) indicates

English tolerates complex margins or consonant clusters in onset position English

therefore ranks the markedness constraint lower than the given faithfulness constraint

One obvious observation that can be made from this analyses is that English loan word

nativization in EkeGusii through anaptyxis leads to an increase of syllables in the resulting

loaned words as is illustrated in (91) repeated from (90) above

91) Number of syllables in the English noun as compared to those in English nouns in

EkeGusii

English noun no of syllables nativised form no of syllables

stͻ lsquostorersquo 1 [sitoo] 2

brʌint lsquobrushrsquo 1 [βurasi] 3

skul lsquoschoolrsquo 1 [sukuru] 3

glɑs lsquoglassrsquo 1 [kerasi] 3

(91) shows that EkeGusii nativized forms have more syllables because EkeGusii is a strict

CV language unlike English This is further illustrated by figure (38) for the English

word skul lsquoschoolrsquo

English EkeGusii

σ σ σ σ

O N C O N O N O N

CC V C C V C V C V

s k u l s u k u r uFigure (38) Increase of number of Syllables in nativized word forms due to anaptyxis

This figure indicates that while the English word form has only one syllable its EkeGusii

nativized form has three This is due to vowel epenthesis

Vowel epenthesis as a strategy of loaned word adaptation in order to satisfy constraints on

phonotactic and syllable structure in the borrowing language is a common feature across

languages It is not unique to EkeGusii language nor indeed to Bantu languages This is

demonstrated by the following examples of nativized loan words into various languages

from English as adapted from Uffimann (2004) in (92)

92) Vowel epenthesis of English loan words in Yoruba Japanese SeTswana Shona

Samoan Sranan

Language example of noun pronunciation gloss

Yoruba kilaasi [kilaasi] class

Japanese sutoraiko [sutoraiko] strike

SeTswana kirisimasi [kirisimasi] christmas

Shona girini [γirini] green

Samoan sikauti [sikauti] scout

Sranan buku [uku] book

In (92) there is at least an epenthetic vowel either to break a complex margincluster or

open a closed syllable For example in SeTswana the word Christmas is nativized to

[kirisimasi] by epenthesizing vowel [i] three times The first epenthesis breaks the

consonant cluster [kr] the second one the cluster [sm] while the third opens the closed

vowel at the end of the loaned word SeTswana like EkeGusii and most other Bantu

languages is a strict CV syllable structure

42212 Nativization of closed syllables

While English sometimes tolerates codas or closed syllables in its syllable structure

because of its nature (has diverse sources of loan words) (Cruttenden 2011) EkeGusii has

a total ban on codas just like most other Bantu languages (Nash 2011 Uffimann 1994)

Thus English syllables with codas entering EkeGusii have to break the codas In this study

this is achievable through paragogic epenthesis which is the insertion of a vowel at the end

of a word syllable (John 2000) as (93) shows

93) Nativization English nouns borrowed into EkeGusii through paragogic epenthesis

fine faɪn baini βainitorch tͻ tint tochi tͻͻtinti machine mǝintin masini masi ni suit sut suti suti(93) shows that all the source forms of the words have closed word end syllables Their

nativized forms however are not instead they are all open as illustrated by figure (39)

for the English word sut lsquosuitrsquo

English EkeGusii

σ σ σ

O V C O N O N

C N C C V C V

s u t s u t i 192

Figure (39) Opening of the English closed syllables by paragogic epenthesis in EkeGusii

This figure shows that the closed English syllable is opened by a paragogic vowel [i] This

like anaptyxis above has the effect of creating an additional syllable in the nativized word

form This in Optimality Theory means that whereas EkeGusii nativized forms obey the

CODA constraint English sometimes does not it instead violates it Thus presupposing

the following constraints NO CODA and DEP IO These two are ranked differently to

analyzes the phenomena in (96) Thus while English ranking is DEP IO gtgt CODA

Ekegusii ranking is CODA gtgt DEP-IO The English word torchrsquo tͻtint and its nativized

form in EkeGusii lsquotochirsquo [tͻͻtinti] are analysed by tableaux (31) and (32) respectively

English input tͻtint lsquotorchrsquo

Constraint ranking DEP IO gtgt CODA

Input tͻtint DEP-IO NOCODA

a) tͻtint

b) tͻtinti

Tableau (432) English realization of the input tͻtint

EkeGusii input [tͻͻtinti]

constraint ranking NO CODA gtgt DEP-IO

Input tͻͻtinti NO CODA DEP IO

a) tͻtint

b) tͻͻtinti

Tableau (433) EkeGusii realization for the input tͻͻtinti

In tableau (32) candidate (a) is the output even though it violates the NO CODA a

markedness constraint This is because this constraint is not of any serious consequence in

determining the output in English DEP-I0 on the other hand if violated by English will

mean that the violating candidate will never be the output The reverse is true in tableau

(33) for EkeGusii NO CODA is of higher rank as compared to DEP ndash IO in EkeGusii

The implication here is that any English syllable with coda entering EkeGusii must be

changed that is nativized to codalessness through vowel paragogic epenthesis in order to

conform to the CV syllable structure of EkeGusii

Complex syllabic margins according to Kager (1999) have a marked status That is why

languages avoid them through vowel epenthesis and consonant deletion processes Some

languages such as Lanakel (Lynch 1974 Blevins 1995) avoid both types of complex

margins (onsets and codas) by vowel epenthesis as (94) adapted from Kager (1999) shows

94) Avoidance of complex margins by vowel epenthesis in Lanakel

Complex margin form epenthesized form gloss

i) t-n-ak-ol t i nagͻl you (sing) will do it

ii) ark-ark argar ikh to growl

iii) k m-n-man-n kamn i man i n for her brother

(94) shows that consonant clusters are not allowed in Lanakel which presupposes the

constraints COPMLEX ONS and COMPLEX CODA In (94i) the vowel [i] breaks an

onset cluster while in (94ii) the vowel breaks a coda cluster This is analyzed in tableau

(34) for the Lanakel input for the word t-n-ak-ol

Input t-n-ak-ol

Constraints COPMLEX ONS gtgt DEP IO (SEG)

Input t-n-ak-ol COPMLEX ONS DEP IO SEG)

a) t-n-ak-ol

b) t i nagͻl

Tableau (434) Lanakel realization of the input [t-n-ak-ol]

Candidate (b) is optimal because it does not allow a complex onset which is banned in the

language Candidate (a) on the other hand loses because it violates the constraint This

phenomenon (vowel epenthesis) is common in the nativization of EkeGusii loanwords from

English as has already been discussed above

According to Batibo (1996) vowel epenthesis involves the insertion of a vowel between

two consonants or after a consonant in a syllable final position The choice of the vowel to

be epenthesized is according to Uffimann determined by a number of factors According

to Pulleyblank (1988) and Shinohara (1997) for example in theie analysis of Yoruba and

Japanese respectively concluded that the maximally unmarked underspecified or

phonetically shortest vowel is inserted via default Paradis (1996) analyzing Fula found that

the stem vowel is copied into the epenthesized vowel slot while Akinalabi (1993) working

on Yoruba Batibo (1995) on SeTswana and Smith (1997) on Sranan concluded that

epenthesis is sometimes controlled by consonant vowel assimilation mostly labial

attraction A combination of all these contributed to the choice of the epenthetic vowel in

this study It does not however mean that it is only vowel sounds that are epenthesized In

some languages such as Axininca Campa (Kager 1999) consonant sounds are also

epenthesized as shown in (95)

95) Consonant epenthesis in Axininca Campa

Underspecified form surface form gloss

no-N-koma-i noŋkomati he will paddle

no-N-koma-ko-i noŋkomatakoti he will paddle for

no-N-tinthik-ako-aa-i-ro noɳtinthikakotaatiro he will cut for it again

Adapted from Kager (1999)

(95) indicates that the underspecified word forms are without consonants in some onset

positions The realizations of these words however are with an epenthesized consonant [t]

This violates the faithfulness constraint DEP-IO (SEG) which demands that output

segments must have input segments This consonant epenthesis is out to break vowel

complex margin COMPLX (V) Axininca Campa consonant epenthesis in (95) is

analysed in tableau (35) for the input no-N-koma-i

Constraint ranking COMPLX (V) gtgt DEP-IO (SEG)

Input no-n-koma-i COMPLX (V) DEP IO (SEG)

a) noŋkomati

b) no-n-koma-i

Tableau (435) Lanakel realization of the input [no-n-koma-i]

Because complex vowel margins are not allowed in Axininca Campa a consonant is

epenthesized to break the complex vowel in candidate (a) the optimal candidate

However there was no case of consonant epenthesis observed in this study

423 Supra-segmental nativization

The third feature that characterize phonological change of EkeGusii loan nouns from

English is Supra segmental or prosodic Supra segmental nativization in this study focused

on toneme structure As was observed in subsection 4141 Ekegusii is a tone language

Katamba (1993) posits that in a tone language pitch can be used to distinguish word

meaning or to convey grammatical distinction It is in this perspective that tone languages

differ from stress (non-tone) ones such as English In other words pitch does not have

these functions in stress languages It is against this background that it can be argued that

EkeGusii loans from English are nativized by tonemization (tone nativization) The English

stressed noun loans are assigned EkeGusii toneme structure they are tonemized as

illustrated by (96)

96) Toneme nativization English nouns in EkeGusii

English noun pronunciation (stressed) EkeGusii nativized form pronunciation (toned)

school skul esukuru [e-s u k u r u]

L H L H

blanket blaeligŋkɪt oboranketi [oβo-r a ŋ k e t i]

L L H L H

chief t int if omochibu [o m o-t int i β u]

L L H L

wire w ǝ ɪ ǝ eguaya [e g w a j a]

L H L H

Adapted from Bickmore (1997)

(96) indicates that while all the English forms are stressed in the initial syllables their

Ekegusii nativized forms are tonemized All the prefixes in the nativized forms are toneless

(low toned) while the initial vowels of the root forms are high toned (Bickmore 1997

Mwita 2012) In other words stress in English is preserved as a variant of high tone in

EkeGusii nativized forms (Kang 2010) Cases of such changes are not unique to EkeGusii

Some other tone languages behave in a similar manner For example in Yoruba spoken in

Nigeria and Shona spoken in Zimbabwe (Kenstowicz 2006) and Dholuo spoken in

Kenya (Owino 2003) loan words from English with stressed syllables are realized with

high tones in these languages as shown in (97)

97) Stress preservation as a variant of high tone in tone languages

a) YorubaEnglish noun pronunciation (stressed) Yoruba nativized form pronunciation (toned) liberty lɪbǝt libati [liHbati]

tomato tǝmatǝʊ tomato [tomaHto]

guarantee gaeligrǝnti garanti [γarantiH]

b) shonaEnglish noun pronunciation (stressed) Shona nativized form pronunciation (toned) Recipe resǝpi resipi [reHsipi]

Philosophy fǝlɒsǝfi firosofi [firoHsofi]

c) Dholuo

English noun pronunciation (stressed) Dholuo nativized form pronunciation (toned) battery baeligtri betiri [beHtiri]

location lǝʊkeɪintn lokesen [lokeHsen]

degree digri digiri [diγiHri]Adapted from Kang (2010)

In (97) the stressed syllables of English are preserved as high tones in the target languages

just like in EkeGusii This study therefore concludes that English stress becomes high tone

in EkeGusii loaned nouns from English

The change of English stress to high tone as observed above can be accounted for within

Optimality Theory specifically by autosegmental OT Bickmore (1997) observes that

within Optimality Theory instead of deriving surface forms from underlying198

representations via the serial application of a number of phonological rules a form is

grammatical if it satisfies a ranked set of constraints better than any other possible

candidate does The candidate set consists of forms created from a given input form by

generator (GEN) the component that generates permutations of the input With respect to

tone it is assumed that GEN can manipulate both tones and their associations to Tone

Bearing Units (TBUs) Thus minimally GEN can add and delete tones themselves as well

as manipulate (that is expand or reduce the size of) input High Tone Spans (HTS)

Following Bickmore (1997) this study analyses EkeGusii loan nouns nativization by

tonemization within the Optimality Theory frameworks (McCarthy and Prince 1995 The

following constraints of these framework are used MAX-IO (F) which demands that no

feature should be deleted DEP-IO which prohibits feature insertion IDENT (F) which

prohibits feature changing between input and output correspondent and UNIFORMITY

which prohibits feature fusing

Following Goldsmith (1990) well formedness conditions for tones that is i) all vowels are

associated with at least one tone ii) all tones are associated with at least one vowel and iii)

association lines do not cross the following constraints ensuring that tones are linked and

that tones bearing units are assigned a tone are used in this study DEP- ET which

prohibits floating of tones (FLOAT) (T) meaning that every tone must have a

correspondent tone bearing unit (TBU) and MAX- IO (T) which specifies that each TBU

must have a correspondent tone (SPECIFY (T)) (Bickmore 1997) Bickmore (1997) ranks

the constraints outlined above as (98)

98) Ekegusii toneme constraints ranking

FLOAT gtgt MAX IO (T) gtgt IDENT (H) gtgt DEP ndash IO (T) gtgt SPEC (T)

In the analysis of loaned words within Optimality Theory the aim is to show that the loans

obey the constraints of the target language and not those of the loaning one This is true

given what has already been indicated EkeGusii is a tone language and therefore can be

analyzed tonally while English is a stress language and cannot be analyzed in terms of tone

The process of tonal nativization of English loans into EkeGusii starts with tonimization

which involves changing the loans from their stress form to a tone form as illustrated in

(97) above It is this tonemized form that is analyzed within OT in this study to account for

EkeGusii loanword nativization by tonemization as in tableau (36) for the

noun ooraŋgeti lsquoblanketrsquo from the input blaeligŋkɪt

Input blaeligŋkɪt

output o β o r a ŋ g e t i

L L H L HThis tone structure presupposes the following tone constraints FLOAT gtgt MAX-IO gtgt

IDENT-IO gtgt DEP-IO

Input oo-raŋget-i

FLOAT MAX-IO

IDENT-IO

DEP-IO SPEC

a) o β o- r a ŋ g e t i

b) o β o- r a ŋ g e t I

c) o β o- r a ŋ g e ti

d) o β o r a ŋ g e t i

Tableau (436) EkeGusii realization of the input oo-raŋget-i Adapted from Bickmore (1997)

The optimal candidate in tableau (35) is (a) because it obeys the highest ranking constraint

FLOAT which prohibits a floating high tone (H) Besides this candidate obeys most of

the constraints violating the relatively lowly ranked SPECIFY (which demands that each

tone bearing unit (TBU) must have a correspondent tone Candidate (b) on the other hand

loses because it violates FLOAT a fatal violation Candidate (c) loses because it changes

a low (L) feature to a H feature This is a serious violation of tone structure in EkeGusii

Candidate (d) deletes the feature (H) suggesting (wrongly) that the language has only L

tones which in essence are underlying

Of significance to this study is the observation that English nouns in EkeGusii adopt the

tone structure of EkeGusii in the process of being accommodated into the phonological

structure of EkeGusii as shown in (97) above

424 Nativization by phonological processes

A phonological process involves a change of a phoneme (segment) in terms of features

which is motivated by a number of factors such as the environment of the phoneme and the

phonemes with which it occurs in a word By environment here it means the position in a

word where a phoneme occurs (Katamba 1993) For example if a voiced English stop

such as d occurs after a voiceless one such as k as in walked it is devoiced as in wͻkt

and not wͻkd and therefore the phonological process involved is that of devoicing This

section gives an analysis of various phonological processes which characterize the

nativization of EkeGusii loaned words from English These are processes which

characterize the phonology of EkeGusii the target language

4241 Nativization by voice dissimilation (Dahlrsquos Law)

According to Bickmore (1998) Dahlrsquos law is a phonological process in which voiceless

consonants in adjacent syllables become dissimilar in some Bantu languages Guthrie

(1967) observes that languages which show the effect of Dahlrsquos law are found within his

zones E20 - E50 ndash F20 and G60 It has also been argued that languages vary a great deal as

to which particular consonants undergo the process which consonants trigger it and how

the process affects multiple targets within the same word (Bennett 1967 and Davy amp

Nurse 1982) According to Bickmore (1998) there is evidence that Dahlrsquos Law affects

EkeGusii as in (99)99) Voice dissimilation (Dahlrsquos law) process in EkeGusii(a) ͻ-kͻ-rͻͻt-a [ͻkͻrͻͻta] lsquoto dreamrsquo o-ko-kor-o [okoγoro] lsquofootrsquo o-ko-nyu-a [okoɳwa] lsquoto drink (b) o-ko-kana [oγokana] lsquoto denyrsquo o-ko-tuua [oγotuua] lsquo to be bluntrsquo ͻkͻsɛka [ͻγͻsɛka] lsquoto laughrsquoSource Bickmore (1998)

In (99a) in the word lsquo[okoγoro]rsquo for example the voiceless velar obstruent (k) in the

prefix oko- demands the voiced velar obstruent (γ) in the root -γor Thus the [ndashvoice]

and [-continuant] features of [k] in the prefix dissimilates to [+voice] and [+continuant] in

[γ] in the root of the word The dissimilation process in this case is from the voiceless

obstruent k underlyingly to the voiced obstruent γ in the surface The reverse is true in

(99b) The [+voice] and [-continuant] underlying features in the prefixes surface as [-voice]

and [+continuant] in the roots

Uffmann (2013) explains this phenomenon using the laryngeal feature [OPEN] The

consonant k which is produced with open larynx dissimilates to the next onset if the

onset also contains a consonant produced in the same manner- open larynx This means that

k will lose its [open] laryngeal feature to become [g] with laryngeal [close] However [g]

is not recognized in EkeGusii phonology except in combination with the nasal

consonant ŋ Therefore as Uffmann suggests in order for the [g] to share its manner

specification with the preceding nasal It loses the [closed] manner feature to become [γ] a

featurally mannerless continuant

The dissimilation processes in (99) can be illustrated by figures (40) and (41) for EkeGusii

words [okoγoro] okogoro lsquolegrsquo and [oγokana] ogokana lsquoto denyrsquo |okokoro| rarr [okoγoro]

o k o k o r o underlying form

o k o γ o r o surface formFigure (40) EkeGusii dissimilation from |k| rarr [γ] |okokana| rarr [oγokana]

o k o k a n a underlying form

o γ o k a n a surface formFigure (40) EkeGusii dissimilation from |γ| rarr [k]Figure (40) shows that k the dorsal stop dissimilates to the dorsal continuant [γ] while it

is the opposite in (41) that is the dorsal continuant dissimilates to the dorsal stop These

dissimilation processes in Optimality Theory suggest the Obligatory Contour Principle203

(OCP) a phonological hypothesis that bans certain consecutive identical features in a

linguistic unit such as a morpheme or word underlyingly (Prince and Smolensky 2004) In

particular dissimilation in EkeGusii seems to target the voiceless dorsal stop feature

leading to a markedness constraint OCP(dorstop) Uffmann (2013) which provides that

there should be no sequence of voiceless dorsal stop and another voiceless stop The

dissimilation also suggests the faithfulness constraint DEP IO (F) which demands that

features of an input segment must be preserved in the output no change Given these

constraints the dissimilation in figure (36) is analyzed in in tableau (37)Input okokoro okogoro lsquolegrsquoConstraints OCP (dorstop) gtgt DEP IO (Seg)

Input okokoro OCP (dorstop) DEP IO (Seg)

a) okokoro

b) okoγoro

Tableau (437) EkeGusii output of the input okokoro

The input form of this tableau loses as candidate (a) because it violates the constraint

OCP (dorstop) a highly ranked constraint in EkeGusii which demands that a voiceless

dorsal stop should not be in a sequence with another voiceless stop Candidate (b) on the

other hand is optimal since it obeys the constraint

Dahlrsquos law seems to affect English loan words into Ekegusii as (100) below shows100) Dahlrsquos law in English nouns borrowed into EkeGusii(a) EkeGusii t s features dissimilate with [g] featuresEnglish noun pronunciation EkeGusii nativized form pronunciation

bank baeligŋk ebenki [e-eŋg-i]

drink driŋk eturunki [e-turuŋg-i]

sink siŋk esinki [e-siŋg-i]

(b) EkeGusii t features dissimilate with [r] features nz features with [t ] features

English noun pronunciation EkeGusii nativized form pronunciationTransport trəensp tכ eturansiboti [e-turanziכt-i]

Transfer traelignsfə eturansiba [e-turanzi-a]

(100) shows disimillation of the feature voice in the nativized word forms suggesting the

OCP (voice) constraint This is illustrated by the EkeGusii nativized forms of the words

[esiŋgi] lsquosinkrsquo and [e-turanzi-a] lsquotransfer respectively in figures (41) and (42)

Oslashs ɪ ŋ k Oslash English Oslash s Oslashŋ kOslash EkeGusii Oslashs i ŋ g Oslash Phonemic substitution

e s i ŋ g i (Nativization Borrowing complete)Figure (41) Substitution of the English k for EkeGusii [g] Oslash d Oslash r ɪ ŋ kOslash English OslashOslash Oslash r Oslashŋk Oslash EkeGusii Oslash t Oslash r u ŋg i Phonemic substitution

e t u r u ŋg i (Nativization Borrowing complete)Figure (42) Substitution of the English d for EkeGusii [t] English k for EkeGusii [g]In figure (41) the EkeGusii k a voiceless velar stop is substituted for [g] a voiced velar

stop This is in order for the two to be dissimilar in the feature [voice] In figure (42) there

are two substitutions of English consonants in EkeGusii First there is substitution of the

English voiced alveolar stop d for EkeGusii voiceless stop [t] which dissimilates the

features [voice] and [continuance] between the substituted [t] and the [r] in the adjacent

syllable Secondly there is substitution of the English k for EkeGusii [g] as in (42)

The dissimilation processes in the nativized words in the given data is not motivated by

prefixes as in the EkeGusii non-nativized words discussed early instead the process seems

to be triggered by the first consonants of the roots In (100a) it is the voiceless obstruents

of the roots ([ts]) that triger the process while in (100b) the initial consonant of the root

is [t] The [t] in the roots of the examples in (100b) is voiceless hence the voiced pre-

consonantal continuant [z] This voiced pre-consonantal continuant triggers voice

dissimilation in the adjacent syllable consonant hence the voiceless continuant [] in the

words [e-turanziכt-i] lsquoeturansibotirsquo (transport) and [e-turanzi-a] lsquoeturansibarsquo

(transfer)

In order to analyze the voice dissimilation shown in (100) for the English loans in EkeGusii

in OT perspectives the OCP(feature) constraint (McCarthy 2004) is presupposed This

constraint bans consecutive identical features in segments of a word The changes that

occur in (100) also suggest the faithfulness constraint IDENT IO (F) which requires that

the features of an input segment be identical in the surface Thus the English and EkeGusii

realizations of the words given in (100) differ only in their ranking of these constraints as

tableaux (38) and (39) for the English input baeligŋk lsquobankrsquo showInput baeligŋkConstraints IDENT (F) gtgt OCP (voice)

Input baeligŋk IDENT (F) OCP (voice)

a) baeligŋk

b) eŋgi

Tableau (438) English output of the input baeligŋk

Output (a) is the winning candidate in the tableau above because it violates the least serious

constraint OCP (voice) in English language Candidate (b) which loses on the other

hand violates the serious constraint in the language IDENT (F) which demands that the

voice features in the input must be maintained in the output This is comparable to tableau

(39) below for EkeGusii realization with reversed constraint ranking from that of EnglishEkeGusii input eeŋgi

Constraint ranking OCP (voice) gtgt IDENT (F)

Input eeŋgi OCP (feature) IDENT (F)

a) eeŋgi b) ebaŋki

Tableau (439) EkeGusii output of the input eeŋgi

Candidate (a) obeys the constraint OCP (voice) which is highly ranked in EkeGusii and

therefore wins while (b) loses because it violates the constraint

Loan word voice dissimilation (Dahlrsquos Law) discussed above is also influenced by the

prefix structure of EkeGusii which is different from that of the loaning language

Basically most nouns in EkeGusii have prefixes which have a bi-morphemic structure The

prefix itself is divisible into two elements that is an initial vowel (augment or pre-prefix)

and the prefix per-se (Bickmore 1998) as (101) demonstrates101) EkeGusii bi-morphemic prefix structure a) o - mo - mur - a b) a - ba - mur- a

augment prefix root FV augment prefix root FV

c) e - ki - ar - a d) e - bi - ar- a

The augment and the prefix in (101) mark the class and the number of the noun For

instance the augment and prefix in (101a) mark the noun as class 1 and that the noun is in

singular while in (101b) the augment and prefix mark the noun as class 2 and that it is in

the plural form

11 boy

finger 8 more than 1

finger

According to Katamba (1989) in many Bantu languages there is a rule which requires that

a consonant in a prefix disagree in voicing with the first consonant of the root it is attached

to a voiced stem initial segment requires a voiceless consonant in the prefix and that a

voiceless stem - initial segment requires a voiced consonant in the prefix The Kirundi

(Burundi) examples in (102) demonstrates this observation

102) Voice dissimilation in Kirundi Imperative 1 st person singular present Word Pronunciation Gloss Word Pronunciation Glossa) rya rjia eat turia tu-rjia I eat mwa mwa shave tumwa tu-mwa I shave va va come from tuva tu-va I come from bona ona seen tubona tu-bona Irsquove seen b) soma soma read dusoma du-soma I read teeka teka cook duteka du-teka I cook seka seka laugh duseka du-seka I laugh kubita kuita hit dukubita du-kuita I hitSource Katamba (1989)(102a) shows that when the prefix is voiced the stem (root) is voiceless the opposite is

true in data (101b) This is Dahlrsquos rule in operation which as illustrated by (100) above

affects English loaned nouns in EkeGusii

Dahlrsquos Law affects English loans into other languages too with similar results as those

discussed in this research For example in Kitharaka English loan words undergo the

process as exemplified in (103)103) Dahlrsquos Law in Kitharaka loaned words from English English word pronunciation Kitharaka nativized form pronunciation

share the feature [- voice] In their surface forms (SF) however the consonants are

dissimilar in terms of voice While those of the roots are voiceless those of the prefixes are

voiced This is Dahlrsquos Law in operation

4242 Nativization by defricativization spirantization (fricativization) and bilabialization

Defricativization as was pointed out in sub-section 41214 is a process where a voiced

continuant or spirant loses its continuant quality by becoming [-CONTINUANT]

Fricativization therefore can be described as the opposite of defricativization In

fricativization or spirantization a [-CONTINUANT] sound becomes [+CONTINUANT]

Bilabialization on the other hand is the process of changing a non-bilabial sound into a

bilabial one These processes characterize EkeGusii loan words from English and are

discussed in 42421 defricativization 42422 fricativizationspirantization and

42423 bilabialization

42421 nativization by defricativization

As pointed out in subsection 41214 above defricativization is a process where a voiced

continuant consonant [+CONTINUANT] loses its continuant feature to become minus

continuant [-CONTINUANT] This process also affects EkeGusii loan words from English

as illustrated in (104)

104) Nativization of English nounsborrowed into EkeGusii through defricativization

a) Nativization of γ as [g]

English noun pronunciation EkeGusii nativized form pronunciation UF SF

tank taeligŋk etanki |etan-γi|rarr[etaŋ-gi]drink drɪŋk eturunki |eturun-γi|rarr[eturuŋ-gi]b) Realization of as [b] English noun pronunciation EkeGusii nativized form pronunciation

camp caeligmp ekambi |ekan-i|rarr[ekam-bi]

remand rɪmaelignd erumande |eruman-ri|rarr[eruman-

c) Realization of r as [d]UF SF

secondary sekǝndri esekenderi |ɛsɛkɛn-ri| rarr [ɛsɛkɛn-dɛri]conductor kɒndʌktǝ ekondagita |ɛk כn-raγita| rarr[ ɛk כn-daγita]

The substitution processes in (104) are further illustrated by figures (43) for the EkeGusii

γ rarr [g] (44) for EkeGusii rarr [b] and (45) for EkeGusii r rarr [d]

English word tank taeligŋk

Nativized from [etaŋgi] lsquoetankirsquo

e t a n γ i EkeGusii underlying (nativized) form

e t a ŋ g i EkeGusii surface realizationFigure (43) Realization of EkeGusii γ as [g]

English word camp kaeligmp

Nativized from [ekambi] lsquoekembirsquo

e k a n i EkeGusii underlying (nativized) form

e k am b i EkeGusii surface realizationFigure (44) Realization of EkeGusii as [b]

English word secondary sekǝndri

Nativized from [esekenderi] lsquoesekenderirsquo

ɛ s ɛ k ɛ n r OslashOslashi EkeGusii underlying (nativized) form

ɛ s ɛ k ɛ n d ɛ r i EkeGusii surface realizationFigure (45) Realization of EkeGusii r as [d]These figures show that an underlying fricative consonant is realized as a stop which is of

the same place of articulation in the surface form of the loan word This is phoneme

defricativization This kind of mapping is analyzed in (105) for the English words taeligŋk

lsquotankrsquo sɪŋk lsquosinkrsquo and drɪŋk lsquodrinkrsquo105) Output mapping English nouns in EkeGusii by defricativization (|γ|rarr[g])

Here the voiced velar fricative γ is realized as [g] a voiced velar stop which is not found

in EkeGusii phonology This change is due to nasal homorganization discussed in section

(41123) The nasal [n] an alveolar is homorganized to [ŋ] by [γ] a velar It is the newly

formed nasal velar [ŋ] a stop which assimilates [γ] a velar fricative to [g] a velar stop

Thus EkeGusii does not have the voiced velar stop [g] it is only realized phonetically

from the voiced fricative γ which is defricativized

The realizations in (104b and c) are analyzed in (106)

106) Nativization of English nouns borrowed into EkeGusii by defricativization (||rarr [b])

i) |e-kaan-i| rarr [ekaambi]

|ɛ-kͻͻniut-al| rarr [ɛkͻmbjuta]

ii) |r| rarr [d]

|e-rumanr-e| rarr [erumande]

|ɛsɛkͻnrari| rarr [ɛsɛkͻndari]

In (106i) the bilabial voiced fricative || underlyingly is defricativized to [b] a voiced

bilabial stop while in (106ii) r an alveolar tap is defricativized to [d] an alveolar stop

The underlying forms in data (105) and (106) above are explained by the fact that EkeGusii

phonemic inventory does not have voiced plosives [g] [d] and [b] instead it has the voiced

fricatives [γ] [r] and [] Therefore the borrowed plosive stops from English are nativized

by defricativization as shown The process of defricativization is such that the plosive is

first fricativized for example b rarr [] before being defricativized in the surface form as

shown in (107)

107) Defricativization of the English noun in EkeGusii after fricativization

camp rarr |kan -- i|rarr [kambi] lsquocamprsquo

Defricativization can be explained by OT using the constraints IDENT IO (F) and

COMPELEX C (because a homorganized nasal consonant NC as has already been

observed is treated as a singletone consonant and not a consonant cluster) (107) above is

analyzed in tableau (40) for illustration

Input |kan-i| gt kambi

Output [kambi]

Constraint ranking COMPLEX C gtgt IDENT IO (F)

Input kani COMPLEX C IDENT IO (F)

a kambi

b kani

c kanbi

Tableau (440) EkeGusii output of the input kani

Candidate (a) is the output in this tableau because it satisfies the highly ranked constraint in

the language COMPLEX C This is because as was observed earlier the combination

nasal consonant (NC) is treated as a pre-consonantal unit (one consonant) and not a

consonant cluster Candidates (b) and (c) violate this constraint they have consonant

clusters (violating COMPLEX C)

Cases of loaned word nativization by defricativization has not been given a lot of attention

This is because as Ohala and Sole (2008) observe defricativization is associated with

nasalized fricatives and that these kind of fricatives are rare or marked and that they tend to

be defricativized if voiced Ohala and Sole further observe that defricativization is as a

result of the difficulty involved in simultaneous production of nasalization and friction

both of which features are marked Defricativization therefore aims at unmarking one of

the features for ease of articulation The target feature in this case being the voiced

fricatives This process characterizes EkeGusii and the English loaned words into EkeGusii

as shown in (107) above

Defricativization though rare as observed by Ohala and Sole (2008) seems to be common

in some Bantu languages many of which do not treat nasal consonants as clusters of

consonants as illustrated in (108) for Kiswahili spoken in East Africa

108) Kiswahili defricativizationWord UF SF Glosskamba |kaN-a| [kamba] rope

ngombe |N-γN-ɛ| [ŋͻmbe] cow

kondoo |kͻ-N-rͻͻ| [kͻndͻͻ] sheep

pingu |piN-γu| [pingu] curffs(ad hoc)

(108) shows that nasals in Kiswahili like in EKeGusii are underspecified in the underlying

form They only receive feature specification on the surface (phonetically) All the

fricatives after the nasals underlying are realized as voiced stops in the surface This is

defricativization

42422 Spirantizationfricativization of the stops

According to Kenstowicz (1994) spirantization is a phonological process which involves a

change of stops to fricatives (spirants) through what phonologists refer to as lenition or

weakening of the stops In this study the process can be said to be motivated by Dahlrsquos

Law of dissimilation discussed in (4241) above as illustrated by (109)

109) Nativization of English nouns by fricativization of bilabial stops

English word pronunciation EkeGusii nativized form pronunciation

dip dɪp tibu |tipu| tiu

pastor pʌstǝ basita |pasita| asita

bolt bͻlt boriti |pͻriti| ͻriti

In these data for example in the English word dɪp lsquodiprsquo which is realized as [tiu] lsquotibu

in EkeGusiirsquo the stop p is realized as [] a voiced fricative This is because the consonant

p must be dissimilar in terms of voice features with the initial root consonant [t] In this

case it becomes [] a [+voice and +continuant] consonant since [t] is [-voice] It becomes

[+ continuant] because EkeGusii does not have the stop p and therefore the bilabial

continuant [] is picked because it is the closest consonant to p in terms of features (they

are both bilabials)

The English velar stop k is equally fricativiced during nativization due to Dahlrsquos Law of

voice dissimilation as exemplified in (110)

110) Nativization of English nouns in EkeGusii by fricativization of velar stops

taxi tᴂksi tegisi |tekisi| [teγisi]

socks sɒks sogisi |sͻkisi| [sͻγisi]

box bɒks bogisi |ͻγisi| [oγisi]

degree dɪgri tigirii |tikirii| [tiγirii]

In (110) the velar k is fricativized into the voiced velar fricative [γ] in the EkeGusii

nativized forms from English This facilitates voice dissimilation of the adjacent syllables

For example in the word [teγisi] lsquotaxirsquo the consonant [γ] of the second syllable is

dissimilar to that of the first syllable [t] at least in terms of voice While [t] is a voiceless

stop [γ] is a voiced continuant

This phenomenon presupposes the markedness constraint OCP(voice) and the faithfulness

constraint IDENT IO(F) as exemplified in tableau (41) for the English word tᴂksi lsquotaxirsquo

Input eteγisiConstraint ranking OCP (voice) gtgt IDENT (F)

Input eteγisi OCP (voice) IDENT (F)

a) eteγisi

b) tᴂksi

Tableau (441) EkeGusii output of the input eteγisi

In this tableau the optimal candidate is (a) because it obeys the constraint which bans

sequences of voiceless dorsal stops and another voiceless stop The dorsal stop k does

therefore not occur with the voiceless stop t consecutively hence the voiced continuant

[γ] as the output

Fricativization or spirantization of stops as discussed above is not a preserve of EkeGusii

Other languages such as Kitharaka (Uffmann 2013) and KiKuria (Oden 1994) among

others show a similar trend as as in (110)

110) Dahlrsquos Law in Kikuria

Word Pronunciation Gloss

okogaamba |okokaamba| [okoγaamba] to say

ogosooka |ͻkͻsͻͻka| [ͻγͻsͻͻka] to respect

ogoteema |ͻkͻtɛɛma| [ͻγͻtɛɛma] to hit

Adapted from Uffmann (2013)

In (110) a prefix stop which is underlyingly voiceless is voiced if the stem or subsequent

syllable starts with a voiceless consonant and the other way round The prefix therefore

dissimilates in voicing from the stem or a following prefix

Dissimilation witnessed in (110) does not involve voice only but frication as well In cases

where there are only stops in adjacent syllables one of stops has to be fricativiced For

example in |ͻkͻtɛɛma| (underlying) there are two stops |k| and |t| therefore one has to be

fricativiced to effect dissimilation In this case it is |k| of the prefix which changes to the

voiced fricative [γ] This is what happens to EkeGusii loaned words as discussed in this

sub-section

42423 Nativization by bilabialization of labiodentals

This like fricativization is a means by which consonants of adjacent syllables get voice

and place features that are dissimilar as in (111)

111) English nounsrrowed into EkeGusii nativization by bilabialization of labiodentalsEnglish noun pronunciation EkeGusii nativized form pronunciation

veranda vǝraeligndǝ baranda -ɸaranda

university junɪvǝsɪtɪ yunibasiti -juniɸasiti

fine faɪn baini -aini

(111) shows that the labiodentals v and f are bilabialized to either the voiceless bilabial

[ɸ] or its voiced counterpart [] depending on the vowel of the syllable which they make If

the vowel is of [Retracted Tongue Root] ([RTR]) as in veranda [ɸaranda] it becomes [-

VOICE] [ɸ] However if it is of [Advanced Tongue Root] ([ATR]) as in tv [tibii] it

becomes [+ voice] []

Bilabialization described like fricativization presupposes the constraints OCP(voice)

and IDENT IO (F) among others The English input vǝrᴂnǝ lsquoverandarsquo in (111) is analyzed

in tableau (42)

EkeGusii input ɸaranda lsquobarandarsquo

Constraint ranking OCP(voice) gtgt IDENT IO (F)

Input ɸaranda OCP(voice) IDENT IO (F)

a ɸaranda

b vǝrᴂndǝ

Tableau (442) EkeGusii output of the input ɸaranda

This tableau shows that candidate (a) is the output because it obeys the constraint

OCP(voice) which is ranked highly in EkeGusii language (b) loses since it disobeys

the constraint in a bid to be true to the input IDENT IO (F) which (a) disobeys is

inconsequential in deciding the output EkeGusii in any case breaking it is motivated by

the voice dissimilation process explained above

4243 Nativization of nasal consonant clusters (NC) by pre-nasalization andhomorganization

In sub- section 41213 it was observed that EkeGusii language does not have obvious

consonant clusters It was argued that any nasal consonant sequence in this language is

treated as a single consonant This it was argued is due to the fact that EkeGusii (a Bantu

language) does not allow consonant clusters or complex margins English on the other hand

allows consonant clusters (section 41131) This means that loaned nouns from English

into EkeGusii with complex margins will have to undergo structural and phonological

changes in order to be accommodated One such modification strategy is that of vowel

epenthesis which breaks the clusters (4221) In this sub-section two segments a nasal

and a consonant will be treated as one a pre-nasalized consonant made by the process of

homorganization as in (112)

112) English nouns with nasal consonant sequence borrowed into EkeGusii

bench bendʒ ebenchi [eβendʒi]

drink drɪŋk turunki [eturuŋgi]

stamp staeligmp esitambu [esitambu

driver draɪvǝ omontereba [omontereɸa]

Each of the nativized words in (112) has a nasal plus consonant sequence Following Herby

and Downing (2005) Clements (1990) and Sievors (1981) this sequence is treated as a

single unit and not a complex consonant or a cluster of consonant in EkeGusii This single

unit (a prenasal consonant) is achievable through pre-nasalization and homorganization

(Cammenga 2002) In this case the consonantal element shares place feature with the nasal

element EkeGusii loaned nouns from English eturunki lsquodrinkrsquo and ebaranda lsquoverandarsquo for

example will be homorganized as in (113)

(113) EkeGusii loan onun homorganization

i) drink

Input e- + -turunγ-i

affixation eturun-ki

prenasalization [eturuŋ-gi]

homorganization [eturuŋgi]

(ii) veranda

Input e-+βaranr-a

affixation eβaran-ra

prenasalization eβaran-da]

homorganization [eβaranda]

In (113i) prenasalization derives the prenasal consonant [ŋg] while in (113ii) the prenasal

consonant derived is [nd] These prenasals are homorganic in that the consonantal219

elements like the nasals have the same place of articulation feature For [ŋg] it is [+velar]

while for [nd] it is (+alveolar) Of significance to be noted here is that the prenasals in the

loan word forms in EkeGusii are treated as a single unit (a prenasal consonant) while they

are treated as a nasal consonant cluster in their English forms as illustrated in figure (47)

English EkeGusii

veranda baranda

vǝrᴂndǝ [βaranda]

lsquoa consonant clusterrsquo lsquoa unitrsquo (one consonant)

Figure (46) English and EkeGusii treatment of nasal + consonant sequences

In optimality theory EkeGusii handling of the sequence in figure (41) presupposes the

markedness constraints COMPLEX C and the faithfulness one FAITH C as analyzed in

tableau (43)

Input ɸaranda (unit)

Constraint ranking COMPLEX gtgt FAITH C

Input ɸaranda (unit) COMPLEX C IDENT IO (labiodental)

a ɸaranda (unit)

b vǝrᴂndǝ (cluster)

Tableau (443) EkeGusii output for the input ɸaranda (unit)

The optimal candidate in this tableau is (a) because it treats the nasal plus consonant

sequence as a single unit and not a cluster As it has already been observed EkeGusii does

not allow consonant clusters unlike in English as (candidate (b) in this tableau

The argument advanced here and elsewhere in this study that there are no nasal plus

consonant clusters in EkeGusii phonology and that the combination nasal plus consonant

is not farfetched In Indonesian language (spoken in Indonesia) (Clements 1983) for

example there is nasal consonant substitution as exemplified by (114)

114) Nasal consonant substitution in Indonesian

UF SF Gloss

a) i mǝN-pilih [mǝmilih] to choose to vote

ii mǝN-tulis [mǝnulis] to write

iii mǝN- kasih [mǝŋasih] to give

b) i mǝN-bǝlih [mǝmbǝlih] to buy

ii mǝN-dapat [mǝndapat to get

iii mǝN-ganti [mǝŋgati] to change

(114) shows that when an input nasal |N| underlyingly is followed by a voiceless obstruents

as in (114a) the obstruent is deleted leaving its place of articulation to the nasal In (114ai)

the nasal is the bilabial [m] in (114aii) the alveolar [n] while in (114aiii) it is the velar [ŋ]

all represented by the archiphoneme |N| underlyingly In data (114b) however the voiced

obstruents that come after the |N| underlyingly are retained in the outputs forming a nasal

plus consonant cluster in Indonesian The observation made about (114b) is that the

underlying nasal represented by the archiphoneme |N| is also realized with the place of

articulation of the following obstruent just as in (114a)

Following these observations therefore it can be generalized that since in nasal plus

voiceless obstruents the obstruent is lost thus remaining with only the nasal which adopts

the place of articulation of the obstruents there is only one consonant in such combinations

even in cases of nasals plus voiced obstruents as in (114b) In (114b) the nasal loses its place

of articulation to the obstruents and joins the obstruents in their places of articulation

Therefore there are no two places of articulation This means that the new sound created by

the two like in (114a) is one and not a cluster of consonant plus nasal

4244 Nativization by vowel height harmonization and disharmonization

As already observed EkeGusii language is characterized by vowel harmony Some of the

loan words depending on their vowel composition undergo vowel harmony Specifically

EkeGusii has mid- vowel perfect harmony and advanced and retracted tongue root pairing

which are vowel height harmonies Similarly EkeGusii is characterized by vowel height

disharmony in which vowels in adjacent syllables in a word do not share height features

(115a and b) gives an illustration of vowel height harmony and disharmony respectively

(115) EkeGusii loan noun nativization by vowel height harmony and disharmonya) Nativization by vowel height harmonyEnglish noun pronunciation EkeGusii nativized form pronunciation

coat kǝʊt egoti [eγoti]

basin baeligsɪn ebeseni [ɛɛsɛni]

lorry lɒri erori [erori]

cheque tintek echeki [tintɛki]

b) vowel height disharmony

tractor trʌktǝ ekeragita [ekeraγita]

feet fɪt ebuti [euti]

break breɪk ebureki [eureki]

(115a) shows vowel height harmony while (115b) vowel height disharmony In [ɛtͻtinti]

lsquotorchrsquo for example the prefix vowel [ɛ] is in height harmony with the root vowel [ͻ]

They are both mid vowels In [ekeraγita] lsquotractorrsquo on the other hand the prefix vowel [e] is

in disharmony with the root vowel [a] while [e] is mid high [a] is low In Optimality

theoretic terms vowel harmony and disharmony as in (115) presuppose the following

faithfulness constraints HARMONY (V) McCarthy (2003) and FAITH V as

demonstrated by analysis the loanword [ɛntɛrɛa] in tableau (44)

EkeGusii input [ɛntɛrɛa]

Constraint ranking HARMONY (V) gtgt FAITH V

Input ɛntɛrɛa HARMONY (V) FAITH V

a ɛntɛrɛa

b draɪvǝ

Tableau (444) EkeGusii output for the input ɛntɛrɛa

In this tableau candidate (a) is optimal while (b) loses This is because candidate (a) does

not violate the highly ranked constraint HARMONY (V) in the EkeGusii while on the

other hand candidate (b) loses because it violates this highly ranked constraint Indeed

EkeGusii unlike English requires that vowels be in harmony

Other languages with vowel harmony in particular vowel height harmony show a similar

trend as in EkeGusii For example in Kera spoken in South West Chad like in most other

languages with vowel height harmony lower vowels and high vowels cannot be in

harmony (Pearce 2003) Pearce observes that when there is a high vowel anywhere in a

word the high feature will spread as illustrated in in (116)116) Kera (Chadic) vowel height harmonyRoot word gloss suffixed form gloss vowelgid- stomach gidi her stomach ici(r)- head ciri her head igud- behind gudi her behind ukas- hand kasa her hand aAdapted from Pearce (2003)In (116) the height of the vowel in the root spreads to the suffix In kas- for example the

low vowel a spreads to the suffix to form kasa in the suffix form This indeed is in

support of EkeGusii occurrences observed in (115a)

Section 42 deals with phonological nativization (segmental phonotactic supra-segmental

changes and phonological processes) that account for the nativization of EkeGusii loans

from English The present section (43) deals with objective three of the study which

analyzes the morphological changes that EkeGusii loan words from English undergo during

nativization It analyzes the morphological processes that account for morphological

adjustments that EkeGusii loaned words from English undergo in order to be

accommodated into the morphological system of EkeGusii The changes are accounted for

within Optimality Theory (Prince and Smolensky 1993 McCarthy 2006)

Morphological change takes place at three levels base word level (root) vowel level and

affix level (Kaspersky 1982) Analyses of morphological change of EkeGusii loaned nouns

from English in this study focus on these levels and are explained by Optimality principles

and guidelines In particular the study shows which morphological ranking is favoured by

EkeGusii outputs given the English loan nouns into EkeGusii as inputs whether that of

English the loaner language or that of EkeGusii the target language This study employs

the following morphological constraints in its analysis

Faithfulness constraints

MAX IO-(MORPH) ndash which demands that there should be no morpheme deletion an input

morpheme should be maintained in the output

MAX- OI (MORPH) ndash this demands that there should be no addition an output morpheme

should have an input correspondent

DEP-IO (MORPH) -no (recipient) affix epenthesis

DEP -IO (V) - no vowel epenthesis

Markedness constraints

COMPLEX (C) ndash no consonant clusters are allowed in syllable margins

ONSET- syllables must have onsets

STRPRES- a structure preservation constraint which demands that the input structure must

be preserved in the output

Alignment constraints

(ALIGN (AFX R RT L))- demands that the right edge of an affix should be aligned to the

left edge of a root an affix should be a prefix

(ALIGN (AFX L RT R))- demands that the left edge of an affix should be aligned to the

right edge of a root an affix should be a suffix

Sources Prince and Smolensky (199394) amp McCarthy (2006)

431 Nativization by nominal classification

As observed in 4121 morphology of the nouns in EkeGusii and English differs in one

crucial manner while EkeGusii nouns are classified into groups known as noun classes

English nouns do not In other words as Demuth (2000) observes Bantu noun classes tend

to be realized as grammatical morphemes rather than independent lexical items that the

classification is part of the larger concordial agreement systems English nouns on the other

hand are characterized as independent lexical items

EkeGusii unlike English enters into a system of pairs of prefixes (morphemes) that mark

the semantico-syntactic (morphosyntactic) categories of singular and plural forms as

demonstrated in (117)

117) EkeGusii noun classification Noun ɛ-nγͻͻkͻ [ɛŋgͻͻkͻ] lsquohenrsquo singular 9a

tinti - nγͻͻkͻ [tintiŋgͻͻkͻ] lsquohensrsquo plural10a

The noun in (134) [nγͻͻkͻ] lsquohenrsquo is in classes 9a in the singular form and 10a in the plural

form respectively Thus (134) shows that all EkeGusii nouns belong to one of the noun

classes identified in the language (Ongarora 2009 Camenga 2002 and Whitely 1965)

The noun classes are marked by the singular and plural prefix markers In (119) for

example the prefix e-n- marks singularity while the prefix tinti-n- marks plurality

In OT theoretic terms the morphology of the noun [n-γͻͻkͻ] in its neutral singular and

plural forms is analyzed in tableaux (45) and (46) as follows

Neutral form analysis

Input nkooko nγͻͻkͻ lsquohenrsquo

The input presupposes the constraints DEP IO (MORPH) - no affix epenthesis and

(ALIGN (AFX R RT L)) (a word must have prefix) The constraints are ranked as

follows

DEP IO (MORPH) gtgt (ALIGN (AFX R RT L))

Input nkooko DEP-IO (MORPH) (ALIGN(AFX R RT L))

a nkooko

b e-nkooko

Tableau (445) EkeGusii neutral form output of the input nkooko

Candidate (a) is optimal in the tableau because it does not violate the serious constraints

DEP-I0 (MORPH) which prohibits epenthesis of an affix to the neutral noun This is

because such affixation will change the meaning of the noun which in its neutral form

means either a lsquohenrsquo or lsquohensrsquo in constructions such as

nkooko teiyo [ŋgͻͻkͻ teijo]-lsquohen not therersquo lsquothere is no hen(s)rsquo

DEP-IO is highly ranked because violating it means that the output will have a totally

different meaning However violating (ALIGN (AFX R RT L)) is not as serious at and

therefore violating it has little significance in determining the output at least in this sense

Analysis of the singular and plural form of EkeGusii word form |e-n-γͻͻkͻ| gt [en-γͻͻkͻ] gt

[eŋgͻͻkͻ] is as follows

Input nkooko nγͻͻkͻ

Constraints and their ranking (ALIGN(AFX R RT L)) gtgt DEP IO (MORPH)

Input nkooko (ALIGN(AFX R RT L)) DEP IO (MORPH)

a e-nkooko

b nk-ooko

Tableau (446) EkeGusii singular output of the input nkooko

Candidate (a) is optimal because it does not violate (ALIGN (AFX R RT L)) a highly

ranked constraint in singular and plural form markings in the language (b) loses because it

violates the constraint Violating it makes the noun neutral it does not refer to any specific

hen Tableau analysis of the plural form is similar to that of the singular form analyzed

above because plural forms like singular ones demand prefixation

English nouns ont the other are not categorized in the same way as the EkeGusii nouns

that is in groups of morphemes paired in singular plural dichotomy determined by

prefixation Instead they are grouped just like nouns from all other languages in terms of

countable versus non-countable concrete versus non-concrete regular versus irregular

nouns among others English Plural and singular forms for the noun lsquoboyrsquo for example can

be analyzed in OT theoretic terms as in tableau (47) and (48) as follows

Input [bͻɪ] ndash singular form

The presupposed constraints here are ranked as follows (ALIGN (AFX L RT R)) gtgt

DEP IO (MORPH) ranked as DEP IO (MORPH) gtgt (ALIGN(AFX L RT R))

Input boy DEP IO (MORPH) (ALIGN(AFX L RT R))

b boy-s

Tableau (447) English realization of the input boy

Candidate (a) is optimal even though it violates (ALIGN (AFX L RT R)) because the

violation is not fatal A singular regular noun in English does not require an affix (suffix)

The output satisfies DEP IO (MORPH) a constraint which is highly ranked in this case

Violating it in this case (and in particular in this word) will be fatal

Opposite ranking as in (47) above determines the output in the plural form of the word as

shown by analysis (48)

Input boy-s- plural form

Constraint ranking (ALIGN (AFX L RT R)) gtgt DEP IO (MORPH)

Input boy-s (ALIGN(AFX L RT R)) DEP IO (MORPH)

a boy-s

Tableau (448) English realization of the input boy-s

The optimal candidate here is (a) It satisfies the requirement in English which requires that

plural forms of regular nouns be affixed with a suffix as embodied in the constraint

(ALIGN (AFX L RT R))

The difference between the noun morphologies of English and EkeGusii analyzed so far

means that EkeGusii loaned words from English undergo a number of changes in order to

be accommodated in EkeGusii morphological structure One such change is for the English

nouns to enter into EkeGusii noun classesgroups as demonstrated in (118)

(118) Nativization of English nouns borrowed into EkeGusii by nominal classification

English noun EkeGusii nativized EkeGusii noun classes

Singular plural form singular plural

scout scouts [sikaoti] 1omo-sikaoti 2aβa-sikaoti

record records [rɛkͻti] 9e-rɛkͻti 10 tinti-rɛkͻti

blanket blankets [raŋgeti] 14 oβo- raŋgeti 6 ama- raŋgeti

pastor pastors [βaasita] 1 omo- βasita 2 aβa-βasita

(118) shows that the English nouns in EkeGusii fall into EkeGusii noun classes in their

nativized forms The word lsquoblanketrsquo for instance enters into classes 14 oβo- and 6

ama- for singular and plural forms respectively This in OT means that the English

forms (input) adopt different structural shapes and therefore violating the faithfulness

constraints FAITH (MORPHEME) (input morphemes are the same no change) MAX OI

(Morpheme) (output morpheme must have an input correspondent no addition and

markedness ones STRPRES (ALIGN (AFX L RT R)) (a suffix) and (ALIGN(AFX R

RT L)) (a prefix) To illustrate this observation the English word blankets lsquoblanketsrsquo

nativized as lsquoama- ranketirsquo is analyzed in tableaux (49) and (50) for English and EkeGusii

outputs respectively

English output blanket-s

Input blanket-s

Constraint ranking STRPRES gtgt (ALIGN (AFX L RT R)) (ALIGN (AFX R RT

Input blanket-s STRPRES (ALIGN (AFX L RT R)) (ALIGN(AFX R RT L))a blanket b ama- ranketi c blanket-s Tableau (449) English output of the input blanket-s

This tableau shows that candidate (c) is the optimal because it preserves the input structure

a constraint which is highly ranked in English concerning plural formation it also obeys

the relatively high ranked constraint in English which demands that plurals be affixed with

a plural marker a suffix Candidate (a) loses because it violates SRTPRES which

disallows a change of structure of the input in output Candidate (b) on the other hand

loses because it prefixes (wrong affixation) instead of suffixing as required by English

besides being in violation of the STRPRES This is compared to EkeGusii output in tableau

(50) as follows

EkeGusii input ama-ranketi

Constraint ranking (ALIGN(AFX R RT L)) gtgt STRPRES (ALIGN (AFX L RT R))

Input ama-ranketi (ALIGN(AFX R RT L)) STRPRES (ALIGN (AFX L RT R))

a blanket

b ama- ranketi

c blanket-s

Tableau (450) EkeGusii output of the input ama-ranketi

In this tableau the alignment constraint which demands prefixation is ranked above the

rest while the constraint which demands for a suffix is ranked least This underscores the

fact that languages rank constraints differently while EkeGusii plural demands a prefix

English demands a suffix

Loan noun nativization by nominal classification is not a preserve of EkeGusii Some other

Bantu languages behave in a similar manner For example in KiNyarwanda (Kagayime

2010) loaned words are allocated to the nominal classes by the Allocation Theory This

kind of allocation depends on either the semantics of the loan noun or its morphology

French loan nouns into KiNyarwanda behave as in (119)

119) French nouns in Kinyarwanda nominal classes

Loan word form French form Class Gloss

u-mu-arabu arabe 1 Arab

a-ba-arabu arabes 2 Arabs

u-mu-note minute 3 minute

i-mi-nota minutes 4 minutes

i--lonji orange 5 orange

a-ma-lonji oranges 6 oranges

i-gi-tari hectare 7 hectare

i-bi-tari hectares 8 hectares

i-katoti carotte 9 carrot

za-karoti carrottes 10 carrots

u-rufanga franc 11 franc

a-ma-fanga francs 6 francs

(119) shows that French nouns enter into Kinyarwanda nominal classes Every French

noun depending on its semantics and morphology joins an appropriate KiNyarwanda

nominal class

Nativization process in KiNyarwanda through nominal classification resembles that of

EkeGusii In both languages the classes into which the various nouns enter are similar and

are determined by the semantics and morphology of the nouns The only difference

between the two nativization processes is that while in EkeGusii the augment vowel is

homorganic to that of the prefix vowel due to vowel harmony which characterizes

EkeGusii phonology in KiNyarwanda on the other hand the augment vowel is in

disharmony with that of the root

Nominal class nativization in KiNyarwanda (Kayigema 2010) is accounted for within the

allocation theory while this research accounts it within Optimality Theory Allocation

theory accounts for the distribution of the loans into nominal classes governed by their

morphology and semantics OT on the other hand accounts choice of nominal classes by

loan words as competition among constraints Allomorphic distribution in both languages

are controlled by Dahlrsquos Law of dissimilation

Languages without Meinhoffrsquos nominal classes do not nativize the same way In other

words nouns in these languages do not recognize nominal classification They therefore

behave differently from those with nominal classes For example in Urdu (Islam 2011)

plurality of noun loans is marked by suffixation like in English as in (120)

120) English noun plural marking in Urdu

English noun plural form Urdu singular form Urdu plural form

plate plate-s plat plat-a

glass glass-es glas glas-a

book book-s buk buk-a

building building-s bilding bilding-a

Adapted from Islam (2011)

(120) shows that plural of the English noun in Urdu is achieved through the suffixation of

the suffix -a ([aelig]) on to the root This is not the case with EkeGusii and most Bantu

languages In Bantu languages plurality is marked by prefixation (and not suffixation) and

it is a function of nominal classes in which it is a singular- plural number pairing of the

As observed in section (4122) choice of nominal class by the nativized forms is not

arbitrary rather it is determined by the semantics of the root word In other words as

Givon (1972) observes noun (stembase) semantics determines prefixes choice For

example the noun -mura lsquoboyrsquo or lsquomale youthrsquo falls within the semantic meaning of

animate human being and in its singular form it takes the prefix omo- class 1 while in

its plural form it takes the prefix aβa- class 2 Similar nouns behave the same way Thus

in (118) above the English words that are borrowed into EkeGusii enter into a specific

class determined by the semantics of the noun rootbase lsquoscoutrsquo for example means an

animate human being and therefore enters class (1) singular form and class (2) plural

form as in (118)

Because semantics plays a major role in determining the membership of noun classes and

their prefixes this study briefly interrogates its role in the process of nativization of the

English loan words into EkeGusii in the following subsection

4311 The role of semantics in morphological nativization of English nouns inEnglish

It has been observed that a noun in EkeGusii and indeed in most Bantus languages with

noun class system (Henderiks amp Poulos 1990 Givon 1972) enters into a specific class

depending on its stem gender number and size (section 4122) In other words class

membership of a noun depends on its meaning and that it is this meaning which determines

the kind of prefix to be affixed on to it The chosen prefix puts the noun in the class it

belongs to ( see table 2 above)

Indeed as the nouns in EkeGusii enter into their classes on the basis of their semantic

content so are the loans from English into the language as further is illustrated in (121)

121) EkeGusii nouns from English and their classesprefixes determined by semantics

Noun class gloss root meaning

omo-gabana 1 governor animate being human

aba-gabana 2 govenors animate being human

obo-ranketi 14 blanket inanimate object

ama-ranketi 6 blankets inanimate objects

eke-ragita 7 tractor inanimate cultural object

ebi-ragita 8 tractors inanimate cultural

objects

e-retio 9 radio inanimate object

chi-retio 10 radios inanimate objects

(121) shows that the English nouns in EkeGusii are affixed with a class marking prefix

which is determined by their root meanings or semantics For example the noun governor

enters class (1) and not any other class because of its semantic features It is [+ANIMATE

+HUMAN BEING] This class demands the prefix omo- The word lsquotractorrsquo on the other

hand falls within the semantic features [-ANIMATE -HUMAN BEING +OBJECT] and

therefore enters its appropriate class- 7eke- Thus the English nouns entering EkeGusii

morphology do not enter haphazardly but rather they are determined by their semantics

That is depending on the meaning of the root of the loan an appropriate class which

preserves the meaning of the input in the output is determined and assigned

This choice of nominal class by loan nouns as a result of their semantic features in

Optimality Theory presupposes Faithfulness constraint which preserve the meaning of the

input in the output form that is MAX IO (meaning) Because EkeGusii nouns must

belong to a noun class and that the noun class is marked by a prefix an appropriate

alignment constraint is also presupposed (ALIGN (AFX R RT L)) which demands that

an affix be a prefix Therefore the loaned word must be prefixed This differs from

affixation of plurality in English which demands for suffix affixation (ALIGN (AFX L

RTR)) (Prince and Smolensky 2004 and McCarthy 2006) Finally the structure of the

English word as input changes in it nativized or output form This means that the structure

preservation constraint (STRPRES) (Golston amp Yang 2001 Aronoff 1998 and Kiparsky

1982) is presupposed This constraint provides that the structure of an input form be

preserved in the output (no change of structure form in the output) These constraints are

ranked differently for English and EkeGusii outputs as analysis in tableaux (51) and (52)

below show

Input tractor-s

Constraints and their ranking MAX IO(meaning) gtgt STRPRES (ALIGN(AFX L RT

R)) (ALIGN(AFX R RT L))

Input tractor-s MAX IO(meaning) STRPRES ALIGN(AFX L RT R) ALIGN(AFX R RT L)a tractor-s b tractor c ebi-ragita

Tableau (451) English output of the input tractor-s

Candidate (a) is the output because it only violates the relatively low ranked constraint in

English which provides that there must be a prefix to mark class and other nominal

features a feature not recognized by English The rest of the other candidates lose because

they violate the highly ranked constraint MAX IO(meaning) for (b) which demands that

the meaning of the input be preserved in the output and STRPRES in (c) which demands

that the structure of the input be preserved in the output This is compared to EkeGusii

analysis

EkeGusii input ebi-ragita

Constraint ranking MAX IO(meaning) gtgt (ALIGN (AFX R RT L)) gtgt (ALIGN (AFX

L RT R)) STRPRES

Input ebi-ragita MAX IO(meaning) (ALIGN(AFXR RT L)) ALIGN(AFX L RT R STRPRES

a tractor-s b eke-ragita cebi-ragita Tableau (452) EkeGusii output of the input ebi-ragita

Candidate (c) is the optimal since it does not violate the constraint demanding that the input

meaning be preserved in outputs This is the determining constraint (b) loses because it

changes the meaning of the input from being in plural to singular Candidate (a) loses

because it aligns the given prefix wrongly in EkeGusii it is a suffix yet EkeGusii demands

a prefix

The analysis of the role of semantics in morphological nativization and OT handling of the

same is one of the major contributions of this research in theoretical linguistics This is

because available literature (Zivenge 2009 Kayigema 2010 amp Islam 2011 among others)

indicate that morphological loan word nativization this far has not focused on the role that

semantics plays in the process None of these studies focuses on the role of semantics in the

process of loanword nativization

(121) above show that the English nouns are pluralized by suffixation (the suffix -s) in

all the given cases However their plurals in EkeGusii nativized forms are prefixed (the

prefixes differ as per the semantics of the noun root) as shown in (122)

122) Pluralization of English and EkeGusii nativized forms

English forms EkeGusii forms class semantics

scouts skaʊts [aβa-siikaouti] 2 animate humam

records rkͻdz [tinti-rɛkɛkͻti] 9 inanimate object

blankets blᴂŋkɪts [ama-raŋgeti] 14 inanimate obj ect

pastors pʌstǝz [aβa-βasita] 2 animate human being

governors gʌvǝnǝz [aβa-γaβana] 2 animate human being

sacraments saeligkrǝmǝmǝnt [ama-sakaramento] 6 inanimate object

These data show that all nouns entering EkeGusii from English are affixed for class and

number This is because each word in EkeGusii belongs to a particular class and number

Given the difference in affixation for plural marking between English words and their

nativized forms in EkeGusii as indicated in (122) above affix location constraints are

presupposed (Prince and Smolensky 2004 and McCarthy 2006) EkeGusii language

demands the following affix location constraint ALLIGN (AFX R RT L) which states

that align the right edge of an affix to the left edge of a root to mark plurality among other

functions To illustrate EkeGusii noun lsquoomotersquo omo-te lsquotreersquo is analyzed in tableau (53)

Input o- mo- te

aug 33PSG tree

output [omote]

This input presupposes the constraints ONSET (ALLIGN (AFX R RT L)) (ALIGN

(AFX L RT R)) ranked as follows ALLIGN (AFX R RT L) gtgt ONSET (ALIGN

(AFX L RT R))c te-omo

Candidate (a) is the output because it does not violate the constraint which requires that the

right edge of an affix be aligned with the left edge of the root to which it is affixed Its

violation of the constraint ONSET and (ALIGN (AFX L RT R)) is of little consequence

in determining the output in EkeGusii Candidate (b) loses because the affix has been

affixed in the wrong part of the root that is to the right edge instead of the left edge as

demanded by the language In essence as McCarthy (2006) observes the affix location

alignment constraint (ALIGN (AFX R RT L)) declares that this affix be a prefix This

kind of affix location alignment affects both the singular and plural forms of EkeGusii The

plural form of the noun [omo-te] is [eme-te] lsquotreesrsquo Its OT analysis will have similar

results as in tableau (53) because the constraints and their ranking are similar The

constraint (ALIGN (AFX R RT L)) which declares that this affix be a prefix means that

the plural marking morpheme be a prefix

However constraint ranking is not the same in English language forms In the plural form

the presupposed constraints will be (ALIGN (AFX L Root R)) which demands that the

left edge of an affix be aligned to the right edge of a root (ALIGN (AFX R RT L)) and

COMPLEX (C) which demands that there should not be a complex or cluster of

consonant at the syllable margins This is illustrated in tableau (54) which analyses the

English word records rekͻds

Input rekͻds lsquorecordsrsquo

This presupposes the following constraints and their ranking (ALIGN (AFX L RT R))

gtgt COMPLEX (C) (ALIGN (AFX R RT L))

Input records (ALIGN (AFX L RT R)) COMPLEX (C) (ALIGN (AFX R RT L))

b record-s

c s-record

Tableau (454) English output of the input record-s

(a) is the output in this tableau because the plural marker affix (which is a suffix in English)

is correctly aligned even if it violates COMPLEX (C) which prohibits consonant clusters

at syllable margins

The singular forms of the English nouns do not require an affix and therefore no affix

location constraint is required The relevant constraint in this case is MAX IO (meaning)

which demands that there should be no change of meaning in the output input meaning

should be maintained This is illustrated by the singular form lsquorecordrsquo rekͻd as analyzed

in tableau (55)

Input recordrekͻ d

This presupposes the following constraints and their ranking MAX-1O (meaning) gtgt

(ALIGN (AFX LR RT LR))

Input rekͻd MAX-1O (meaning) (ALIGN (AFX LR RT LR))

a rekͻd-s

b s-rekͻd

c rekͻd

Tableau (455) English output of the input record

Candidates (a) and (b) lose in the tableau because they are affixed affixation and

prefixation respectively These affixations are banned in singular forms of English which

ranks them highly in the language and which demands that a singular form should not be

affixed with any morpheme in English (ALIGN (AFX LR RT LR)) is satisfied in (a)

and (b) because the candidates are affixed as required by the constraint (a) aligns the left

edge of an affix to the right edge of a root - a suffix) while (b) aligns the right edge of an

affix to the right edge of a root- a prefix) However this satisfaction is inconsequential

because the constraint is relatively low ranked in the language regarding singular forms in

the grammar of English

The data indicate that all the English nouns borrowed by EkeGusii have to be nativized

that is they have to enter into a given a noun class These classes as has been observed in

this subsection are marked by prefixation The English noun loans into EkeGusii are

therefore prefixed in order to be admitted into the various EkeGusii noun classes The

constraint which demands for this prefixation as has been observed is (ALIGN (AFX R

RT L)) that is align the right edge of an affix to the left edge of a root To illustrate

EkeGusii loaned word [ɛrɛkͻti] lsquorecordrsquo is analyzed in tableu (56)

Input e-rekoti rekͻd lsquorecordrsquo

e-rekoti [ ɛ- rɛkͻti ]

933PSG- record

This means that the word has been nativized into class 9 and that it is in the third person

singular This presupposes the following constraints and their ranking (ALIGN (AFX R

RT L)) gtgt DEP IO (MORPH)

Input e-rekoti (ALIGN (AFX LR ROOT LR)) DEP IO (MORPH)

a e-rekoti

b record

c record-s

Tableau (456) EkeGusii output of the singular input e-rekoti242

Candidate (a) is optimal because it does not violate the alignment constraint which is

highly ranked in EkeGusii Violating it is fatal because the given word will not be prefixed

for class and therefore will not be classified The loaned word in tableau (54) above has

been effectively prefixed and nativized into class 9 marked by the prefix e- or ɛ-

The right edge of the prefix ɛ- is correctly aligned to the left edge of the root -rɛkͻti as

demanded by the constraint The constraint DEP IO (MORPH) is of no consequence here

though it is of great significance in determining English outputs where it is relatively

highly ranked

EkeGusii plural form of the word ɛrɛkͻti lsquorecordrsquo behaves in a similar manner in terms

of affixation only that changing it to plural will change it in nominal class and number as

illustrated in tableau (55) below

Input records rekͻds

Output chi-rekoti [tinti- rɛkͻti] lsquorecordsrsquo

103PPL record

Here the noun is in class 10 and in plural The constraints pre-supposed are the same as

those used in the analysis of the singular form in tableau (56) above (ALIGN (AFX R

Input chi-rekoti (ALIGN (AFX LR RT LR)) DEP IO (MORPH)

a chi-rekoti

b record

c record-chi

Tableau (457) EkeGusii output of the input chi-rekoti

Candidate (a) is optimal since it violates the less serious constraint in the tableau DEP IO

(MORPH) The rest of the candidates violate the serious constraint and therefore are fatal

violations

Most of the borrowed nouns into EkeGusii from English it is observed seem to favour

certain classes over others In fact majority of the borrowed nouns enter classes 9 (e-) 10

chi- 9 (a) e-n- and 10 (a) tinti-n- a few enter classes 1 omo- 2 aβa- and 6 ama- and

in rare cases into other classes such as 14 oβo- as in obo-ranketi lsquoblanketrsquo The rest

of the classes do not seem to be favoured at all This is because most of the borrowed nouns

name newly invented things objects and names of places (institutions) and that these

nouns belong to the mentioned classes Kayigema (2010)

The following subsections give a description and analysis of how EkeGusii loan words

from English are nativized into EkeGusii nominal classes The nominal classes on focus

are 12 34 14 6 7 8 and 12

4312 Nativization into classes 1 and 2 Nouns

Classes 1 and 2 nouns refer to human beings with class 1 denoting the singular form of

the noun while class 2 denotes the plural In EkeGusii the augment in the singular form is

o- while in the plural it is a- as in o-mo-nto lsquopersonrsquo and a-ba-nto lsquopersonsrsquo

respectively The nominal prefix on the other hand is mo- in the singular form and ba-

in the plural form Loaned words from English entered these classes in EkeGusii as (123)

123) Nativization of English nouns borrowed into EkeGusii in noun classes 1 and 2

English noun EkeGusii form nativized form (class 1) nativized (form class 2)

chief chiibu o-mo-chiibu [o-m-tinti-iβu] a-ba-chiibu [aβa-tintiiβu]

governor gaabana o-mo-gabaana [omo-γa-aɸana] a-ba-gabana[aβa-γaaɸana]

pastor baasita o-mo-baasita [omo-βa-asita] a-ba-baasita [aβa-βaasita]

councilor kansara o-mo-kansara [omo-ka-anzara] a-ba-kansara [aβa-kaanzara ]

All the English nouns borrowed into EkeGusii in (123) have common semantic features

they refer to animate beings (specifically human beings) Upon entering EkeGusii

language the English words are morphologically nativized as has already been observed

In their singular forms they are prefixed with the singular and person marker prefix

omo- of nominal class 1 while in their plural form the prefix changes to aβa- of

nominal class 2 which marks plurality and person This is unlike in their English forms

where in the singular form it is not affixed at all while in the plural it is suffixed As has

been observed in this research these affixations in EkeGusii and the non-affixation in the

English singular form presuppose the OT markedness constraint (Align (Afx R root L))

which demands that the affixes have to be prefixed which is in violation of faithfulness

constraints such as ONSET and DEP IO (MORPH) which prohibit onsetless syllables and

epenthesis of an affix (morpheme) respectively These arguments are captured in tableaux

(58) and (59) for EkeGusii and English outputs for the English inputs chief and chief-

Input o-mo-chiibu lsquoomochiibursquo

Constraint ranking (Align (Afx R root L)) gtgt DEP IO (MORPH) ONSET

Input 0-mo-chiibu (ALIGN (AFX LR RT LR)) DEP IO (MORPH) ONSET

a o-mo- chiibu

b chief

c chief-s

Tableau (458) EkeGusii output of the English input omo-chiibu

In this tableau the optimal candidate is (a) because it obeys the alignment constraint which

is highly ranked in EkeGusii (b) is not optimal because it disobeys the alignment constraint

which leads to a change of meaning of the input from being singular to being neutral (c)

loses because it does not only affix a nonexistent affix in EkeGusii but also a wrong

alignment a suffix instead of a prefix The plural form lsquoa-ba-chiibursquo chiefs will have a

similar analysis This is compared to the English realization in tableau (59)

Input chief

Constraint ranking DEP IO (MORPH) ONSET gtgt (Align (Afx R root L))

Input chief DEP IO (MORPH) ONSET (ALIGN (AFX LR RT LR))

a o-mo- chiibu

b chief

c chief-s

Tableau (459) English output of the input chief

Candidate (b) is the output in this tableau because it is faithful to the input as required by

the constraint DEP IO (MORPH) which is the most highly ranked of the given set of

constraints It requires that the singular forms of the input be maintained in shape in their

outputs The plural form of English unlike that of EkeGusii requires an alignment

constraint which demands for a suffix (English plural is marked by suffixation) and not a

prefix

4313 Nativization of English loans into EkeGusii in classes 1b mo- and 2b [ba-

These classes have the characteristics of classes 1 and 2 only that they lack the augment

element as shown in the loan words (124)

124) Nativization of English nouns borrowed into EkeGusii in noun classes 1Oslash and 2Oslash

Source noun nativized form 1bOslash- 2bOslash-

chief |tinti-ɸu| [motintiβu] [ɸatintiiβu]

pastor |ɸasita| [moɸasita] [ɸaɸaasita]

councilor |kansara| [mokanzara] [ɸakaanzara]

father |ɸaata| [moɸaata] [ɸaɸaata]

In all the cases in (124) both the singular and the plural forms of the nativized nouns are

marked by Oslash- in both classes As Cammenga (2002) observes these are non-augmented

forms which are acceptable in the language under certain circumstances as in [tintiiβu taijͻ]

lsquochief(s) is not therersquo in a case where somebody was checking if there is a chief(s) present

Here the root may carry the meaning of plural or singular Therefore form classes

considered here are instances of lexically determined allomorphy Cases of zero ([Oslash-])

prefixation as Cammenga observes are rare

These classes according to Kayigema (2010) denote to things like trees ditches rivers

natural phenomena and some parts of the body Class 3 denote singular forms while 4

denote plurals

Only one word was collected into these classes as (125) shows nativization of loans into

classes 3 and 4

125) Nativization of English nouns borrowed into EkeGusii noun classes 3 and 4

Word nativized form class 3 class 4

motor car tͻkaa [tͻkaa] ͻmͻ-tͻkaa ɛmɛ-tͻkaa

Nativization in these classes were found to be like nativization in classes 1 and 2 above

As observed in section 41227 these classes are marked by a combination of

corresponding singular and plural prefixes as in (143) below repeated from section

126) EkeGusii noun classes 5 6 7812 and 14 prefixes

Singular plural

(a) 7 eke - 8 eβi-

12 aka - 8 eβi-

12 aka - 14 oβo-

(b) 5 eri- 6 ama-

As observed already whenever a word belonging to some other class is transferred to any

of the classes in (126a) at least the idea of diminution is necessary added to its basic

meaning Whenever some such word is transferred to the class in (126b) at least the idea of

augmentation is added to its basic meaning Loan words too behave the same way Words

from other classes transferred into the classes in (126a) above get the idea of diminution

and those entering (126b) get the idea of augmentation as demonstrated in (127)

127) Nativization of English nouns in EkeGusii by diminution and augmentation

i) Source word nativized form classes

grease griz -ris-i 7eke-risi [ekerisi] 8 [eɸi-risi] lsquoneutralrsquo

school skʊl -sukuur-u 7eke-sukuuru[eγe-sukuuru]8[eβisukuuru] lsquosmallrsquoschools

12 aka-sukuru [aγasukuru] 8[oβosukuru] lsquosmall schoolsrsquo

skirt skɜt -sikaat-i 7eke-sikaati [eγesikaati] 8 [eβisikaati] lsquosmmal skirtrsquo

12aka-sikaati [aγasikaati] 8 [eβi-sikaati]

room rum -rum-u 12aka-ruumu [akaruumu] 14 [oβoruumu] lsquosmall roomsrsquo

7eke-ruumu [ekeruumu] 8 [eβiruumu]

ii) Source word nativized form classes

torch tͻtint tͻtinti 5 rii-tͻͻtiint [riitͻͻtinti] 6[amatͻͻtinti] lsquobig torchesrsquo

governor gʌvǝnǝ kaɸana 5rii-kaɸaana [riiγaβaana] 6[amaγabaana]lsquobig governorsrsquo

(127) shows that loaned words belonging to a given class when transferred to any of the

classes identified in (127i) will be deminutivized For example the word lsquoskirtrsquo is

borrowed into classes 9 e- in singular and 10 tinti- in plural [e-sikaati] and [tinti-sikaati]

respectively However as data (127i) show the word can be transferred into classes 7 and

8 and get the meaning of diminution 7 [aγa-sikaati] lsquoa small skirtrsquo 8 [eβi-sikaati] lsquosmall

skirtsrsquo

Words borrowed into other classes and then transferred to classes 5 (rii-) and 6 (ama-) get

augmented as in (127ii) For example the word lsquogovernorrsquo is nativized into classes 1 [omo-

γaβaana] for singular and 2 [aβa- γaβaana] for plural respectively When transferred to

classes 5 (rii-) and 6 (ama-) it gets the meaning of augmentation (big in stature) lsquoa big

governorrsquo (which may be pejorative or non-pejorative) in class 5 and big governors in

class 6

Other Bantu languages such as KiNyarwanda (Kayigema 2010) and Tong (Zivenge

2009) treat loaned words in a similar manner In other words the loan words into these

languages are nativized through nominal classification For example in KiNyarwanda

(spoken in Rwanda) French words into it are nativized as in (128)

128) French noun nativization in Kinyarwanda

French noun Kinyarwanda morphological form nominal class gloss

chauffeur u-mu-shoferi 1 driver

chauffeurs a-ba-shoferi 2 drivers

meacutedaille u-mu-dari 3 medal

meacutedailles i-mi-dari 4 medals

coat i-kotiri-koti 5 coat

coat a-ma-koti 6 coats

quinine i-kinini 7 tablet

quinine ibi-kinini 7 tablets

(128) shows that French loans into KiNyarwanda like those of EkeGusii are allocated

particular noun classes dependent on the semantic features of the noun The noun

lsquochauffeurrsquo (driver) for example enters class 1 for singular and 2 for plural because these

are classes reserved for the semantic features [+animate +human] The noun lsquomeacutedaillersquo

(medal) on the other hand is allocated classes 3 and 4 because it is characterized by the

features [+inanimate -human +object]

432 Prefixation and pre-prefixation

The previous section has analyzed how EkeG English nouns in EkeGusii are nativized in

the various nominal classes present in EkeGusii One of the main determinants of these

classes as was observed is affixation The prefix it was observed determines whether a

noun belongs to class 1 2 or 3 among others The prefixes on the other hand are

determined by the semantics of the roots of the loan nouns In this section the nature of the

prefix determining noun classes of the loan words and how the loan words from English are

prefixed in order to be accommodated into the morphological structure of EkeGusii are

analyzed

It has been observed that EkeGusii has two types of prefixes that is the prefix per-se and

the pre-prefix (augment) (Cammenga 2002 Ongarora 2008 and Whiteley 1965) Section

(4321) below deals with the prefix while (4322) analyzes the pre-prefix

4321 Nativization by prefixation

Cammenga (2002) observes that EkeGusii roots are regularly prefixed by at least one of the

morpho-syntactic class prefixes The kind of prefix affixed on to a root depends on the

semantic content of the noun root (Giacutevon 1972 Henderikse and Poulos 1990) This is

illustrated in (129)

(129) Nativization of English nouns borrowed into EkeGusii by prefixation

English noun EkeGusii form prefix class prefixed form

carrot kᴂrǝt -karati 9 e- eka-rati

blanket blᴂŋkɪt -raŋgti 14 bo - bo-raŋgeti

ticket tɪkɪt -tikɛti 9 e- e-tikɛti

cabbage kᴂbɪdʒ -kaβitinti 9 e- eka-βitinti

(129) shows that whenever an English noun enters EkeGusii morphology it undergoes

class prefixation in order to be accommodated The prefix chosen by a noun is not

arbitrary it is determined by the semantics of the noun Katamba (1993) observes that

nouns in Bantu are grouped into classes often on a minimally semantic basis which is

dependent on what the nouns refer to whether humananimate or on the basis of other

important properties denoted by the noun For example the English noun lsquocarrotrsquo falls into

class prefixes 9e- for singular and 10tinti- for plural Classes 9 and 10 prefixes

accommodate nouns within the semantic content of animals people body parts tools

instruments household effects natural phenomena among others A large number of nouns

are accommodated within these semantic content classes This explains why most of the

borrowed words fall into the classes

The prefix has CV syllable structures except that of class (9) which has a syllable structure

of V as in (146) above In OT theoretic terms this prefix structure presupposes dominance

of the markedness constraint ONSET over (Align (AFX R RT L)) ranked as ONSET gtgt

(ALIGN (AFX R RT L)) The nativized form of the word lsquoblanketrsquo (146) can be

analyzed in tableau (60)

Input bo-ranketi

Input bo-ranketi (ALIGN (Afx R Root L)) DEP IO (MORPH)

a bo- ranketi

b ranketi

c blanket

Tableau (460) EkeGusii output of the English input bo-ranketi

Candidate (a) is the output because it does not violate the serious constraints Violating

DEP IO (MORPH) is not as fatal as violating the alignment constraint (ALIGN (AFX R

RT L)) demands that loans be affixed with prefixes

The shape of the allomorphs of EkeGusii like many other Bantu languages Kikuyu

(Mwihaki 1998) Kitharaka (Mberia 2004) KiKamba (Mutua 2007) and KiNyarwanda

(Kayigema 2010) among others) prefixes are determined by Dahrsquol Law of voice

dissimilation discussed at length in section 4241 above

Some other Bantu languages also nativize loaned nouns by prefixation For example

Tonga regularly prefixes loaned noun to mark class just like in EkeGusii as illustrated by

130) Tonga English noun prefixation

English Tonga morphological form prefix function

apostle mu-positoli mu- class 1 marker

apostles va-positoli va- class 2 marker

machine mu-china mu- class 3 marker

machines mi-china mi- class 4 marker

girl ri-gelu ri- class 5 marker

girls ma-gelu ma- class 6 marker

school chi-kolo chi- class 7 marker

schools zvi-kolo zvi- class 8 marker

Source Zivenge (2009)

(130) shows that Tonga like EkeGusii and most other Bantu languages nativize noun

loans by prefixation to allocate them appropriate nominal classes determined by the

semantics of the given noun For example the noun lsquoapostlersquo positoli in Tonga is prefixed

with mu- which carries the semantic features [+animate +human] of class one Its plural

form va- marks class 2 The difference between Tonga and EkeGusii as data (130)

shows is that while the Tonga prefix is a strictCV- syllable form EkeGusii prefix allows

an augment and therefore is a (V)CV- form (130) also shows that unlike EkeGusii

prefix (and quite uncharacteristically of Bantu phonology) the Tonga nominal prefix

(zvi-) has a cluster of consonants or a complex margin

4322 Nativization by pre-prefixation

An augment (pre-prefix) is the vowel that is affixed to the prefix in Bantu lexical items

(Kayigema 2010) According to Kayigema common nouns of all kinds allow an augment

while proper nouns and other nouns denoting kinship terms places among others do not

take the augment Different languages use different vowels as augments depending on a

number of phonological factors such as whether a language is characterized by vowel

harmony or not KinNyarwanda for example utilizes four vowels or augments u- o- i-

a- while EkeGusii has only three as illustrated by (131)

131) EkeGusii augments

Augment nominal classes prefixed

a- 2612 as in a-ba-nto lsquopeoplersquo a-ma-riso lsquoeyesrsquo a-ka-gaakarsquosmall old manrsquo

e- 45789 as in e-mete(tree) e-riso(eye) e-geita(gate) e-bi-ita(gates)esese (dog)

o- 131415 as in o-mo-onto (person) o-bo-koombe (hoe) o-ko-gooro (leg)

The rest of the remaining nominal prefix classes (10 11 16 and 21) do not take augments

Cammenga (2002) observes that in most instances the full EkeGusii prefix properly

consists of an augmented prefix that is the classical Bantu combination of an augment

vowel v- also called pre-prefix or initial vowel with a prefix proper usually consisting

of a consonant plus a vowel cv- as discussed in section 42221 This means that

underlyingly the representation of a full EkeGusii prefix has the form v-cv- According

to Cammenga this form covers nominal prefixes in at least classes 1-8 and 10-15 Indeed

as Cammenga (2002) and Kayigema (2010) observe the presence or absence of the

augment is determined lexically by lexical category membership or lexically determined

allomorphy Common nouns of all types are normally pre-prefixed across Bantu languages

Nouns denoting proper names kinship terms and places among others on the other hand

are not augmented as illustrated by Kinyarwanda examples in (132)

132) ( i) KiNyarwanda nominal augmentation

Noun morphological form nominal class gloss

umuntu u-mu-ntu 1 person

abantu a-ba-ntu 2 persons

umuserebenya u-mu-serebenya 3 lizard

imiserebenya i-mi-serebenya 4 lizards

(ii) KiNyarwanda non-augmentation

Noun nominal class gloss

-data 1a Oslash- my father

-nyogukuru 1a Oslash- my grand mother

-Kigali 1aOslash- name of a place

(Kigali)

- Kivu 1aOslash- name of a lake (Kivu)

Adapted from Kayigema (2010)

In (132i) the nominal category of common nouns which allow augmentation (132ii) on the

other hand gives a category of nouns that denote kinship and place and therefore do not

allow augmentation This is in support of EkeGusii morphological behavior regarding

augmentation

With respect to EkeGusii nouns (the focus of this study) lexical category membership

determines that nominal prefixes in morphosyntactic classes 1-8 and 11-15 may or may not

be augmented whereas the other prefixes 1b Oslash- 9 (a) e- (n) 10 (a) tinti-(n) 16a-

and 21ɳ- are not augmented (Cammenga 2002 Kayigema 2010) It therefore means

that borrowed words from English into EkeGusii that fall within the morphosyntactic

classes 1-8 and 11-15 may or may not be augmented as illustrated by (133)

133) prefixation with augmentation of English nouns borrowed into EkeGusii

Word morphological form class and number gloss

i) omogabana [omoγaβana] o-mo-gaban-a 1 SG governor

ii) abagabana [aβaγaβana] a-ba-gaban-a 2 PL governors

iii) risiti [risiti] ri-sit-i 5SG receipt256

iv) amarisiti [amariisiti] a-ma-risit-i 6PL receipts

v) ekereti |ekereti| [egereti] e-ge-ret-i 7SG crate

vi) ebireti [eβireti] e-bi-ret-i 8PL crates

vii) agasukuru |akasukuru| [aγasukuru] a-ga-sukur-u 12SG micro school

viii)ebisukuru [eβisukuru] e-bi-sukur-u 8PL micro schools

ix) oboranketi [oβoraŋgeti] o-bo-ranket-i 14SG blanket

x) amaranketi [amaraŋketi] a-ma-ranket-i 6PL blankets

The following observations are made about (133) Firstly with the exception of classes 1

and 2 very few loans are admitted into the rest of the classes In fact some of the classes

(3 4 and 15) did not admit any while classes (5) and (14) admitted one loan each

Secondly the prefixes in each class have two elements initial vowel (the augment) and the

prefix per se In all the cases however the augment is not compulsory It may or may not

be there though its absence leads to a difference in meaning as was observed in section

41212 For example (133i and ii) above can do away with the augment as in (134)

134) prefixation without augment

i) mogabana [moγaβana] mogabana 1SG governorii) bagabana [βaγaβana] ba- gabana 2PL governor

(134) shows that a prefix can do without an augment and still carry the gender and number

features of the noun it is attached to Pre-prefixation in (134) above presupposes the

constraints ONSET (syllables must have onsets) and DEP V (which prohibits vowel

epenthesis either prothesis or anaptyxis) ONSET in this particular case is ranked higher

than DEP V that is DEPV is dominated by ONSET Thus ONSET gtgt DEP V This is

analyzed in tableau (61)257

Input gabana

Input gaβana DEP IO (V) ONSET

a) o-mo-gaβana

b) mo-gaβana

c) gaβana

Tableau (461) EkeGusii output of the input gaβana

This tableau shows that candidate (a) is the winner The candidate wins because it obeys

the constraint DEP V which is ranked higher than ONSET EkeGusii prohibits onsets

especially in nouns that refer to particular number and gender

CHAPTER FIVE

SUMMARY CONCLUSIONS AND RECOMMENDATIONS

51 Summary

This study investigates the nature of phonological and morphological features and

processes that characterize nativization of English Nouns borrowed into EkeGusii in order

to understand the phonology and morphology of EkeGusii The study examines how

English nouns borrowed into EkeGusii are adjusted phonologically and morphologically in

EkeGusii phonological and morphological environments in order to be accommodated The

study targeted the phonological and morphological processes and features that account for

the differences between English and EkeGusii phonology and morphology Thus the

selected phonological and morphological processes and features were those that enabled

the observation and accounting for the phonological and morphological changes that affect

English nouns entering EkeGusii linguistic environment This was achieved through a

step-by-step procedural exploration of the objectives of the study in Chapter four This

chapter gives a summary of the findings conclusions recommendations and suggestions

for further research

The first objective of the study describes the phonological and morphological structures of

the two languages under investigation- EkeGusii and English Phonologically findings

indicate that the vowel systems of the two languages differ While EkeGusii has a total of

fourteen pure or monophthong vowels English has twenty-five twelve monophthongs

eight diphthongs and five triphthongs It was also established that the acoustic nature of

EkeGusii vowels as produced by native speakers of the language differs significantly from

that of English

Another finding is that the two languages under investigation have some consonants found

in both while other consonants are found in only one of the languages and not the other

This is one of the main contributions of this study

Some phonological processes were found to affect EkeGusii noun phonology and not the

English noun phonology These include feature dissimilation prenasalization

homorganization declusterization of nasal consonants and consonant glides

defricativization and nasal re-syllabifiation

Phonotactically the study established that the syllable structures of the two languages are

different in that while EkeGusii is a strict (V)CV language English is a (C) V (C)

language in which case the consonants can be in cluster forms depending on the word in

question It was further observed that English unlike EkeGusii allows clusters of

consonants of up to three in the onset and four in the coda positions of the syllable

Prosodically findings of this study established that while EkeGusii is a tone language

English is a stress language

Morphologically the findings established that EkeGusii nouns like those in most other

Bantu languages are grouped into morphosyntactic class systems in which the classes tend

to be realized as grammatical morphemes rather than independent lexical items This is not

the case with the English noun which is realized as an independent lexical item in most

cases EkeGusii noun morphology like the morphology of other Bantu languages it was

further established is characterized by a pre-prefix (augment) This is not the case in

English

The second objective of the study analyzed the phonological changes that English nouns

borrowed into EkeGusii undergo during nativization Analyses were carried out within

Optimality Theory

Segmentally the findings of the study show that English sounds not present in EkeGusii

are substituted for those present in EkeGusii phonology The substitution involves

phonemic and feature change which is consistent with OT which provides that languages

rank constraints differently depending on their grammar and that it is this ranking which is

responsible for language differences

Findings further established that English diphthongs and triphthongs are monophthongized

in EkeGusii This process is also consistent with OTrsquos argument that puts markedness

constraints in conflict with faithfulness constraints Thus the markedness constraint

COMPLEX (V) is in conflict with the faithfulness constraint DEP IO (FEATURE) in this

case Analyses indicated that EkeGusii ranks COMPLEX (V) higher than DEP IO

(FEATURE) while the opposite is true in English

Phonotactically findings of the study show that EkeGusii unlike English is a strict (C)V

language All the foreign syllable structures of English were re-syllabified in EkeGusii to

conform to its syllable structure English syllables with complex margins were changed by

epenthesis which broke the complex margins and opened the syllables in EkeGusii OTrsquos

explanation of this observation is that while English allows consonant clusters and

complex onsets and codas EkeGusii disallows them Thus EkeGusii ranks COMPLEX

(C) constraint higher as compared to English

Analysis further show that closed syllables from English into EkeGusii were opened

through epenthesis (paragogic) In OT this is explained by a number of constraints such

as CODA MAX IO (SEG) and IDENT IO Findings established thatCODA is ranked

higher in EkeGusii as compared to English while MAX OI(SEG) ranks high in English as

compared to EkeGusii

Suprasegmentally the findings established that English nouns with stressed forms entering

EkeGusii are tonemized Thus the change of the feature stress in English to the feature

tone in EkeGusii is as explained in OT by the constraints IDENT IO (FEATURE) and

SPEC (T) which demands that each tone bearing unit (TBU) must have a corresponding

tone It was established that while English prefers stress by ranking IDENT (FEATURE)-

STRESS highly as opposed to SPEC (T) EkeGusii does the opposite

Findings further show that phonologically a number of processes characterize nativization

Such processes include voice dissimilation phoneme fricativization or spirantization

phoneme defricativization phoneme bilabialization and vowel harmony and disharmony

Analyses indicate that these processes are governed by EkeGusii constraint ranking For

example consonants in the English nouns in EkeGusii undergo voice dissimilation This

process is determined by Optimality Theory markedness constraint OCP (VOICE) which

is ranked over the faithfulness constraints such as IDENT IO (FEATURE) This finding is

one of the major contributions of this study since no known study has ever targeted voice

dissimilation in loan words

Another finding that is of significance to this study is that nasal consonant clusters from

English lose their cluster status through prenasalization and hormorganization Treating

these combinations as single units in EkeGusii is supported by OT markedness constraint

which bans complex vowels COMPLEX (C)

The third objective of the study analyzes the morphological changes that English nouns

borrowed into EkeGusii undergo before they are accommodated in the morphological

system of the language Analyses focused on nominal classification prefixation and pre-

prefixation

Findings show that English nouns enter the nominal classes (a characteristic of EkeGusii

morphology) and that nominal classification is determined by the semantic features of the

borrowed noun This characteristic in Optimality Theory is explained by principles which

preserve input meaning in the outputs such as MAX IO (meaning) principles dealing with

affix alignment such as (ALIGN (AFX RRT L)) and (ALIGN(AFX LRT R)) and

principles which preserve structure such as (STRPRES) The finding that semantics plays a

major role in morphological nativization is another major contribution of this study to

theoretical linguistics because it shed light on the role of semantics in nativization

Analysis indicate that affixation processes in the languages under study differ Optimality

Theory handles affixation using Affix Alignment Principles among other constraints

Findings show that during morphological nativization the English plural marking suffix lsquo-

s is dropped and prefixes used in its place in the nativized forms The prefixes used to

mark EkeGusii plurals are determined by the class to which the noun in question belongs

Therefore English nouns in EkeGusii are prefixed for plural variously This feature

besides being governed by alignment constraints of OT is also explained by featural

markedness constraints such as OCP (VOICE) and VTV among others This occurrence

makes contribution to theoretical linguistics because it sheds light on the role of affixation

in nativization

52 Conclusions

Based on the findings of this study the following conclusions are drawn Firstly the

phonological and morphological systems of EkeGusii and English are significantly

different Phonologically the phonemic phonotactic prosodic and phonological processes

between the languages are different while morphologically noun classification systems

and affixation processes differ quite significantly between the two languages

Of significance to note are the phonological findings that EkeGusii and English vowel

segments differ acoustically as spectrographic analyses show English stress is tonemized

in EkeGusii and EkeGusii phonological processes not present in English such as voiced

stops fricativization and defricativization vowel harmonization and disharmonization

feature dissimilation and declusterization of nasal plus consonant clusters characterize

English nouns in EkeGusii

It is also worth noting that morphologically the semantics of the stems of English nouns

borrowed into EkeGusii determine the nominal class into which the nouns enter the bi-

morphemic structure of EkeGusii prefix characterize English nouns borrowed into

EkeGusii and affixations in the English nouns obey that of EkeGusii in which plurality

and singularity are prefixed and class paired

Another conclusion is that the phenomenon of noun nativization in EkeGusii can be

accounted for within Optimality Theory a constraint- based approach Through this theory

an explanation to the phonological and morphological adjustments of English nouns

borrowed into EkeGusii is possible Phonologically the main strategies employed in the

nativization of English nouns in EkeGusii are motivated by markedness constraints such as

OCP (V) CODA and COMPLEX which dominate the faithfulness constraints such as

IDENT IO and MAX IO The opposite is true in the analysis of English noun This

observation lead to the conclusion that English allows marked constraints as compared to

EkeGusii Morphologically alignment constraints such as (ALIGN (AFX RRT L))

which outrank faithfulness constraints such as STRPRES motivate nativization of English

nouns in EkeGusii Thus phonological and morphological nativization of English nouns in

EkeGusii is motivated by EkeGusii ranking of the universal linguistic constraints proposed

in OT Therefore ranking of constraints in EkeGusii is responsible for the outputs of

EkeGusii English nouns in EkeGusii This rules out any possibility that the target language

has influence in the phonology and morphology of the target language besides the lexical

item itself (Owino 2003)

53 Recommendations

In the description of EkeGusii vowels a spectrographic acoustic analysis of the vowels was

attempted Many areas of EkeGusii phonetics and phonology such as consonant segments

pitch tone and intensity among others require such an especially in these areas

Therefore this study recommends spectrographic (computer software) analyses of all the

aspects of EkeGusii phonetics and phonology for better understanding and documentation

of the language This study describes the phonological and morphological structures of EkeGusii as a basis

for analyzing English nouns in EkeGusii In the analysis of the morphology of the English

nouns it was established that semantics plays a major role in determining the noun classes

into which the English noun enters a major observation yet this is not given the attention it

deserves in this study or anywhere else It is therefore recommended that a study focusing

on the same be conducted in an effort to shed more light into EkeGusii loan words

nativization

A number of phonological processes were found to characterize nativization of English

nouns in EkeGusii These processes were not given the attention they deserve given the

broad nature of this study It is therefore recommended that a study specifically focusing on

all the phonological processes characterizing English nouns in EkeGusii including the ones

identified in this study be carried out so as to shed more light into the phenomenon of loan

word nativization

EkeGusii has borrowed heavily from a number of languages most notably Kiswahili

English and Dholuo Yet this study has focused only on English It is therefore

recommended that other studies be directed to these languages too if a comprehensive

inventory of all the EkeGusii loan words is to be made This is because there is a likelihood

that EkeGusii indigenous words are facing extinction due to this borrowing and

nativization of these foreign words

Analyses in this study focused on the noun class But other classes especially the verb

which is rich in morphology are equally borrowed and nativized Therefore a study on

these other classes is recommended as it would shed more light on EkeGusii loaned words

nativization especially morphologically

As much as this study has provided important data and advanced illuminating discussions

there were a number of interesting areas that remain unexplored This study it is hoped

will stimulate further inquiry into the areas of EkeGusii orthography syntax and

nativization so as to deepen the phonological and morphological understanding of

linguistic integration Since phonological and morphological systems of a language are

important for the development of the orthographies of a language it is also hoped that other

such studies will stem from the current one since EkeGusii is a language without a

comprehensive orthography Developing orthography for a language ensures the

languagersquos continued existence and its assertiveness as an independent language that can

handle loans fully

The theoretical framework designed for this research is a constraint based generative one

(Optimality Theory) and findings of the study are best explained by such a paradigm Other

researches may also emerge testing the same language phenomena but taking other

linguistic theoretical paradigms to enhance understanding of English loans in the EkeGusii

linguistic environment from a number of theoretical approaches

The study is also hoped to be used as a basis to further constructive studies relating to

Bantu languages other than EkeGusii Since EkeGusii is a Bantu language the findings

from this study can be an lsquoeye openerrsquo and insightful to the understanding of similar

languages in a diglossic situation with English

REFERENCES

Abdul M (2004) The effects of language contact the case of Arab and Igbo

(Unpublished PhD dissertation Khartoum University of Africa)Aikhenvald A Y (2000) Classifiers a typology of noun categorization devices

Oxford OUPAlderete N (1999) Morphologically governed accenti in Optimality Theory

(Unpublished PhD thesis University of Massachusetts Amherst)Anyona M (2011) Phonological influence of EkeGusii in the pronunciation of English (Unpublished MA thesis Egerton University)Appel C Muysken P (1981) Halfway between Quenchua and Spanish the case of

reflexivization In Highfield A and Valdman A(eds) Historicity and variation

in Creole studies Michigan Ann ArborArbib M (Ed) (2006) Handbook of brain theory and neural networks (2nd edn) Mas

MIT pressArchangeli D and Pulleyblank D (2007) Harmony In The Cambridge Handbook of

phonology Paul de Lacy(ed) 353-378 Cambridge Cambridge University press Archangeli D and Pulleyblank D (1992) Grounded phonology Cambridge MIT

pressArchangeli D (1997) Optimality Theory an introduction to linguistics in the 1990s

Optimality Theory an overview (ed) By Diana Archangeli and D Terence

Langendeon1-32 Oxford BlackwelAsher R and Simpson J (1994) (eds) The encyclopedia of language and linguistics

Oxford Pergamon press Ltd RWBailey K (1978) Methods of social research (3rd ed)New York Free PressBarker M (1969) The phonological adaptation of French loanwords into Vietnamese

Mon- khmer studies Journal 3 138-147

Batibo H (1996) Loanword clusters nativization rules in Tswana and Swahili a

comparative study South African Journal of African languages Vol 16 2 33-41Belvins J (1974) Lanakel Phonology (Phd dissertation University of Hawaii)Bennett P (1967) Dahlrsquos Law and Thagisu African language studies8127-159Bickmore L (1998) Metathesis and Dahlrsquos law in EkeGusii Studies in linguistic

sciences Vol28Bicmore L (1998) Opacity effects of Optimal Domains Theoryevidence from EkeGusii

In Maddieson I and Hinnebusch T (eds) Trends in African linguistics 2

Language and HistoryBickmore L (1991) Compensatory lengthening in Kinyambo Katamba I (ed)

Lacustrine Bantu Phonology Afrikanistiche Arbeitspapiere 2575-103Bickmore L (1997) Problems in constructing high tone spared in EkeGusii Lingua

1024 265-290Block B and Tragger G (1972) Outlines of linguistic analysis Linguistic society of

AmericaBloomfield L (1933) Language New York Holt Reihart ampWinstonBoersma P and Weenink D (1992) Doing phonetics by computer Version 6021 SIL

EncoreTM Summer Institute of LinguisticsBoersman P and Hamann V (2009) Loanword adaptation as first language

phonological perception In Calabrese and Wetzels p 11-58Booij G and Rabach J (1984) Morphology and prosodic domains in lexical phonology

In Veira S (ed) Yearbook Vrije Vrije University Vol1 Number1 88-102Bosha I (1993) The Influrnce of Arabic language on Kiswahili with a trilingual

dictionary (Swahili- Arabic- English) Der es Salaam Der es Salaam University

PressBrown F (2004) Adaptation of English words in Scandinavian languages In Colonel P

(ed) Linguistic fieldwork Oslo University of Oslo 22-40Bosire B (1993) Ekegusii a comparative study of Rogoro and Maate dialects (MA thesis University of Nairobi)Bosire K and Machogu G (2013) Authoritative EkeGusii dictionaryendabaro

endabasiayrsquoEkeGusii Nairobi EkeGusii encyclopedia project

Bresnan J and Mchombo S (1987) Topic pronoun and agreement in Chichewa

language 63(4) 741-782 Bright J (1970) Teaching English as a second language London Longman group Ltd Broselow E (2006) Loanword phonology In Keith Brown (ed) Encyclopedia of

language and linguistics vol 7 2nd edn 286ndash290 Oxford Elsevier Broselow E (2009) Stress adaptation in loanword phonology Perception and

learnability In Paul Boersma amp Silke Hamann (eds) Phonology in perception

191ndash234 Berlin amp New York Mouton de Gruyter Brown C (2000) The interrelation between speech perception and phonological

acquisition from infant to adult In John Archibald (ed) Second language

acquisition and linguistic theory 4ndash65 Oxford Blackwell Burenhult N (2001) Loanword phonology in Jahai Lund University Department of

Linguistics Working Papers 48 5ndash14Broselow E (1999) Stress epenthesis and segment transformation in Selayarese loans

Proceedings of the Annual Meeting Berkeley Linguistics Society 25 311ndash325Bynon T (1977) Historical linguistics Cambridge CUPCalabrese A (2009) Perception production and acoustic inputs in loanword phonology

In Calabrese and Wetzels (2009) 59ndash114 Calabrese A and Wetzels L (eds) 2009 Loan phonology Amsterdam and

Philadelphia John Benjamins Cammenge J (2002) Phonology and Morphology of Ekeusii Postfach Rudiga Koppe

VerlagCampbell G (1991) Compendium Of the wordrsquos Languages Vol2 LondonRoutledgeChang C (2009) English loanword adaptation in Burmese Journal of the Southeast

Asian Linguistics Society 1 77ndash94Carsteins V (1991) The morphology and syntax of determiners phrase in Kiswahili

(Unpublished PhD dissertation UCLA)Carsteins V (1993) On the morphology and DP structure In Sam A Mchombo (ed)

Theoretical aspects of Bantu grammar Stanford CSLI publicationsChange C (2009) English loanword adaptation into Burmese Journal of the South east

Asian linguistics society 177-94Chebanne A and Phili C (2015) The phonologization of English loanwords in Kalanga

Marang Journal of language and literature vol26

Chimhundu H (1979) Some problems relating to incorporation of loanwords in the

lexicon (PhD dissertation University of Zimbabwe) Dissertation Abstracts

International Vol 11 Number 1 75-91

Chomsky N and Halle M (1968 1991) The sound patterns of English New York

Harper and Row

Clark J Yallop C and Fletcher J (2007) An introduction to phonetics and phonology

Malden Blakwell Publishing

Clements G (2001) Representational economy in constraint based phonology In Alan

Hall(ed) Distincive feature theory 71-146 Berlin amp New York Mounton de

Gruyter

Clements G (1986) Compensatory lengthening and consonant germination in Luganda

Leo Wetzels and Engin Sezer (ed) Studies in compensatory lengthening

Dordrecht Foris Laboratory Phonology I

Clements G (1976) Vowel harmony in non linear generative phonology Havard MS

Clements G and Keyser S (1983) CV Phonology a generative phonology theory of the

syllable Cambridge Mass MIT Press

Collins B and Mees I (2013) Practical phonetics and phonology Oxon Routledge

Crysal D (1987) The Cambridge encyclopedia of Language Cambridge CUP

David O (2005) Introducing phonology New York Cambridge University press

Davy J and Nurse D (1982) Synchronic versions of Dahlrsquos Law the multiple

application of a phonological dissimilation rule Journal of African languages and

linguistics 4157-195

Davidson L (2007) The relationship between the perception of non-native phonotactics

and loanword adaptation Phonology 24 261ndash286

Davidson L and Noyer R (1997) Loanword phonology in Huave Nativization and the

ranking of faithfulness constraints Proceedings of the west coast conference on

formal linguistics 15 65-79

De Lacy P (ed) (2007) The Cambridge handbook of phonology Cambridge CUP

Demuth K (2000) Bantu noun class systems Loanward and acquisition evidence of

sematic proclivity In G senft (ed) classification systems Cambridge University

press PP 270-292

Denesi M (1985) Canadian Italian a case in point of how language adapts to

environment Italians in Ontario Multicultural history society of Ontario

Doupoux E amp peperkamp S (2002) The phonetic filter hypothesishow phonology

impacts speech perception (and vice versa) Paper presented at the second

International Conference on contrast in pholnlogy University of Toronto May 3-5

2002 Retrieved April 10 2015 from httpwww chss

Utorontocacontrastdupouxpepakamp ppt1

Eichhoff J (1990) Aspects of German borrowing into American English In Nelde P

(Ed) Language in contact and conflict Weshbaden Franz stainer Verlag aMBH

Elwell R (2005) A morphosyntactic analysis of the EkeGusii verb (Unpublished

Honoursrsquo thesis University of Albany)

Ferguson C A (1963) Assumptions about nasals a sample study in phonological

universals In J Greenberg (ed) Universals of language 53-60 Cambridge MA

MIT press

Gay L (1981) Educational research compentence for analysis and application Charcles

E Mairil publishing company London

Givon T (1972) Studies in Chibemba and Bantu grammar In Studies in African

linguistics Supplement 3

Gleason H A (1961) An introduction to descriptive linguistics New Delhi Oxford

GOK (2009) Kenya Bureau of statistics Government printers

Goldsmith J (1990) Autosegmental and metrical phonology Oxford Basil Blackwel

Good J (2005) Reconstructing morpheme order in Bantu the case of causativization and

applicativization Dia chronica 223-57

Golston C and Yang P (2001) White Hmong loanword phonology In Caroline Fery

Antony Dubach Green amp Ruben Van de Vijver(eds) Proceedings of HILP5 40-57

Grimes B (1996) Ethnologue Languages of the World Dallas Texas summer Institute

of Linguistics

Grosjean F (1982) Life with Two Languages An Interoduction to Bilingualism

Cambridge MA Harvard University Press

Gussenhoven C and Jacobson H (2013) Understanding phonology New York

Routledge

Guthrie G (1967-71) Comparative Bantu Farnborough Gregg Vols 1-4

Hadebe S (2002) The standardization of Ndebele languagethrough dictionary masking

Oslo University of Oslo

Hafez O (1996) Phonological and morphological integration of loanwords into Egytian

Arabic EgyptyMonde Arabic premiere serie 27-28 383-410

Hall T and Hamann S (2003) Loanword Nativization in German Journal of Zeitschrift

fur sprachwissenchaft 221 (2003) 56-85

Hancin-Bhatt B (1994) Segment transfer a consequence of a dynamicsystem Second

language research 10241-261

Harjulla L (2005) The Ha noun class system revisted MS

Haspelmath M and Sims A (2002)Understanding morphology London Hodder

Education

Haugen E (1950) The analysis of linguistic borrowing Language 26 210ndash231

Herd J (2005) Loanword adaptation and the evaluation of similarity Toronto working

papers in linguistics 2465-116

Haugen E (1950) The analysis of linguistic borrowing Language 26 210-231

Hock H (1991) Principles of historical linguistics New York Mounton de Gruyter

Hocket C (1970) A course in modern linguistics New York MacMillan

Hoffer B (2005) Language borrowing and diffusion an overview In International

communication studies XL200 Retrieved from http www uri Edu iaics content

2002vlln40120Bates20L 20 Hoffer Pdf

Holden K (1976) Assimilation rates of borrowings and phonological productivity

Language 52 131ndash147

Hooper J (1976) An introduction to Natural Generative Phonology New York

Academic Press

Hsieh F ampand Kenstowicz M (2008) Phonetic knowledge in tonal adaptation

Mandarin and English loanwords in Lhasa Tibetan Journal of East Asian

Linguistics 17 279ndash297

Hussain Q (2011) Vowel substitution a comparative study of English loans in Punjabi

and Urdu International Journal of linguistics 31 E31

Hyman L (1985) A theory of phonological weight Dordrecht Foris

Hyman L (1979) Aghem grammatical structure SCOPL No7 Califonia UCLA

Hyman L (1970) The role borrowing in the justification of phonological grammars

Studies in African linguistics1 1-48

Hyman L and Byarushengo E (1984) A model of Haya tonology In Clements G and

Goldsmith J (eds) Autosegmental studies in Bantu tone Dordrecht Foris

Islam R A (2011) The morphology of loanwords in Urdu the Persian Arabic and

English strands( Unpublished PhD thesis Newcastle university)

Ito C and Kenstowicz M (2009) Mandarin loanwords in Yanbian Korean I Laryngeal

features Onin Kenkyuu [Phonological Studies] 12 61ndash72

Ito C Kang Y and Kenstowicz M (2006) The adaptation of Japanese loanwords into

Korean MIT Working Papers in Linguistics 52 65ndash104

Itocirc J and Mester A (1995) Japanese phonology In John Goldsmith (ed) The

handbook of phonological theory817-838 Cambridge MA and Oxford Blackwel

Itocirc J and Mester A (1999) The phonological lexicon In Natsuko Tsujimura (ed) The

handbook of Japanese linguistics 62-100 Malden MA and Oxford Blackwell

Itocirc J and Mester A (2001) Covert generalizations in Optimality Theory the role of

stratalfaithfulness constraints Studies in phonetics and morphology 7 273-299

Jacobs H and Gussenhoven C (2000) Loan phonology perception salience the

lexicon and OT In Joost Dekkers Frank van der leeuw and Jeron vande Weijer

(eds) Optimalitytheory phonology syntx and acquisition pp193-210

OxfordOUP

Jakobson R (1962) Selected writings I phonologicalstudies Gravenhage Mouton

John A (2000) The Handbook of phonology theory Amazon Blackwell publishers

Jones D (1956) The pronunciationof English Cambridge CUP

John H (1994) Borrowing In John H (ed) Yearbook Vol 2 1 Harmondsworth

University of Warsaw 122-140

Johnson K (1977) Acoustic and auditory phonetics Cambridge CUP

Kadenge M and Mudzingwa B (2012) Comparing Chishona loanwords of monolingual

and bilingual speaker an Optimality Theory analysis South African Journal of

African languages 2012 32 (20 141-151)

Kadenge M (2003) The influence of the Shona phonology on the English speech of First

Language (L1) Shona Speakers of English-With Specific Reference to Vowels

(Honours Dissertation University of Zimbabwe)

Kager R (1999) Optimality Theory Cambridge Cambridge University Press

Kang H (1996) English loanwords in Korean Studies in Phonetics Phonology and

Morphology 2 21ndash47

Kang Y (2010) Tutorial overview suprasegmental adaptation in loanwords Lingua 120

2295-2310

Kang Y (2011) Loanword phonology In Mark van Oostendorp Colin Ewen Elizabeth

Hume amp Karen Rice (eds) Companion to phonology Wiley Blackwell

Katamba F (1989) An introduction to phonology London Longman Academic Press

Katamba F (1993) Morphology Palgrave Hampshire

Katayama M (1998) Optimality Theory and Japanese loanword phonology (PhD

dissertation University of Califonia Santa Cruz)

Kay G (1996) English loanwords in Japanese World Englishes Vol 14 1 67-76

Kayigema L (2010) Loanword allocation in Kinyarwanda (UnpuplishedMA thesis

University of South Africa South Africa)

Kenstowicz M (2005) The phonetics and phonology of loanword adaptation In S- J

Rhee (ed) Proceedings of ECKL Proceedings of 1st European conference of

Korean linguistics 17-32Hankook publishing co

Kenstowicz M and Shuchato A (2006) Issues in loanword adaptation a case study

from Thai Lingua 116 921-949

Kenstowicz M and Sohn H (2001) Accentual adaptation in North kyungsang Korean

In Michael Kenstowicz (ed) Ken Hale A life in language 239-270 Campridge

MA MIT Press

Kihm A (2005) Whatrsquos is in a noun Noun classes gender and nounness MS

Kim H (2008) Loanword adaptation between Japanese and Korean Evidence for L1

feature-driven perception Journal of East Asian Linguistics 17 331ndash346

Kim Y (2008) Topics in the phonology and of San Francisco del mar Huave

(Unpublished PhD dissertation University of Califonia Berkeley)

Kiparsky P and Mohanan S (1982) From Cyclic Phonology to Lexical Phonology In

van de Hulst H and Smith N (eds) The structure of phonological

representations New Delly Indiana University Press Vol 1 1 Number 1 130-

Kombo D and Tromp L (2006) Proposal and Thesis writing an introduction Nairobi

Paulines publiations Africa

Kumar R (1999) Research Methodology a step-by-step guide for beginners London

SAGE Publications Ltd

LaCharite D and Paradis C (2005) Category preservation and proximity versus

phonetic approximation in loanword adaptation Linguistic inquiry 36 223-258

Ladefoged P and Keith J (2001) A course in Phonetics Boston Cernage learning Inc

Langacker R (1968) Language and its structure some fundamental linguistics concepts

New York Holt Reihart amp Winston

Laparombara H (2013) An Optimality Theory account of phonological variation in

Kimbrlek Krioi (Unpublished senior thesis Yalle University)

Libert B (1971) Linguistics and the new language teacher New York MacMillan

Lodhi A (2000) Oriental influences in Swahili a study in languages and culture

contacts

Lombard L (ed) (2001) Segmental phonologyin Optimality Theory constraints and

representation Cambridge amp New York CUP

Massamba D (1991) Phonological theory History and development Dar es salaam Dar

es salaam University press

Massaro D and Michael C (1983) Phonological context in speech perception

Perception and psychophysics 34 338-348

Mathangwana J (1999) Ikalanga phonetics and phonology a synchronic and diachronic

study Califonia CSLI Publishers

Mberia K (2004) The phonology of borrowed words in Kitharaka In Buregeya A

Schroeder H and Eshun F (eds) Occasional papers in language and linguistics

(PP45-57) Department of Linguistics and African languages University of Nairobi

McCarthy A C (2002) An introduction to English morphology Edinburgh Edniburgh

University Press

McCarthy J and Prince A (1993) Prosodic morphology constraint interaction and

satisfaction London Routledge

McCarthy J (2001) A thematic guide to Optimality Theory Cambridge Cambridge

University press

McCarthy J (ed) (2003) Optimality Theory in phonology a reader Malden MA and

Oxford Blacwel

McCarthy J (2003) Phonological processes assimilation In Frawley J (ed)

International encyclopedia of linguistics (2nd edn) 320-323 Oxford University

McCarthy J (2006) Morphology optimality Theory Amherst Elsevier Ltd

McCarthy J and Prince A (1994) The emergence of the unmarked optimality in

prosodic morphology London Equinox

Mchombo S (1967) Bantu grammatical reconstructions Tervuern Tervuren

Mchombo S (2004) The syntax of Chichewa Cambridge CUP

Mecha G (2006) The phonology and morphology of EkeGusii reduplication an

optimality theoretic approach (Unpublished masters thesis Kenyatta university)

Mecha G (2013) Information structure in EkeGusii (Unpublished PhD thesis University

of Nairobi 2013)

Meeussen A (1993) Comparative Bantu test cases for method African language studies

14 6-18

Miao R (2005) Loanword adaptation in Mandarin Chinese perceptual phonological

and sociological factors (unpublished PhD dissertation Stony brook University

Mihalicek V Wilson C (Eds) (2011) Language files (11nth edn) The Ohio

University Press Columbus

Mugenda O and Mugenda A (1999) Research methods quantitative and qualitative

approaches Nairobi ACTS Press

Mutua B (2007) A constraint based analysis of Kikamba nativized loanwords

(Unpublished masters thesis Kenyatta University)

Mwihaki A (2001) Consonantndashvowel harmony Evidence from the phonotactics of

loanword adaptation Poznan Studies in Contemporary Linguistics 37 139ndash145

Mwihaki A (1998) Loanword nativization a generative view of phonological adaptation

of Gikuyu loanwords ( Unpublished PhD thesis Kenyatta University)

Mwita L (2012) Noun tonology in Kuria International Journal of Humanities and social

sciences 2(8) 159-168

Mwita L (2009) The adaptation of Swahili loanwords from Arabic a constraint based

analysis The journal of pan African studies vol2 no8

Nash C (2011) Tone in EkeGusii a constrint based description of nominal and verbal

tonology (Unpublished PhD thesis University of Califonia Santa Barbar)

Nasukawa K (2004) Word final consonants arguments against a coda analysis

Proceedings of the 58th conference of the Tohoku English literacy society 47-53

Nurse D and Philemon G (2006) Commontense tense aspect markers in Bantu in

journal of African languages and linguistics

Nurse D and Philemon G (2006) The Bantu languages London Routledge

Orsquo Connor J (1967) Better English pronunciation Cambridge CUP

Oden D (2005) Introducing Phonology Cambridge Cambridge University press

Oden D (1996) The phonology and morphology of Kimatuumbi Oxford Clarendon

Ogechi N (2011) (ed) Themes in language education amp development in Kenya Nairobi

Nsemia Inc Publishers

Ohala J and Sole M (2008) Turbulence and phnology UC Berkeley phonology lab

annual report

Ohso M (1971) A phonological study of some English loanwords from in Japanese

(Unpublished MA thesis the Ohio State University)

Ongarora D (2009) Bantu morphosyntax a study of EkeGusii (Unpublished PhD thesis

Jawaharlal Nehru University India)

Ongarora D (1996) Vowel harmony in the Rogoro dialect of Ekegusii ( Unpublished MA

thesis Egerton University)

Osinde K (1988) Ekegusii phonology an analysis of the major consonants (MA thesis

University of Nairobi)

Owino D (2003) Phonological nativization of Dholuo loanwords (Doctoral

dissertation University of Pretoria)

Paradis C and Lacharite D (2009) English loanword in Old Quebec French Fewewer

bilinguals does not mean a great increase in native phonetic approximation

Langues et linguistique 32 82-117

Paradis C (1996) The inadequacy of filters and faithfulness in loanword adaptation In

Jacques Durand and Benard Laks (eds) Current trends in phonology modes and

methods pp509-534 Salford University of Saloford Publication

Paradis C and Lacharite D (1997) Preservation and minimality in loanword adaptation

Journal of linguistics 33 pp 379-430

Pearce M (2003) Vowel harmony in Kera (Chadic) (MA dissertation University College

London)

Peperkamp S and Dupoux E (2003) Reinterpreting loanword adaptations the role of

perception In Sole M RecasensD and Romero J (eds) procedings of the 15th

international congress of phonetic sciences 367-370 Barcelona Causal

productions

Peperkamp S and Dupoux E (2001) Loanword adaptations laboratoire de sciences

cognitives et psycolinguitique Universite de Paris 8 httpwww Ehess

FrcentresLSCPpersons dupouxloanwords pdf

Plag I (2002) Morphological productivity structural constraints in English Derivation

Berlin Routledge

Plag I (2003) Word formation in English Cambridge CUP

Prince A and Smolensky P (19932004) Prosodic morphology constraint ineraction

and satisfaction in generative grammar Oxford Blacwell

Pulleyblank D (1988) Vocalic underspecification in Yoruba Linguistic Inquiry 19 233-

Quirk R and Greenbaum J (1990) A university grammar of English England

Longman

Raiz A (2011) The morphology of loanwords in Urdu the Persian Arabic and English

strands (PhD dissertation Newcastle University)

Roach P (1983) English phonetics and phonologyCabridge CUP

Robins R (1964) General linguistics An introductory survey London Longman

Rodgers J (1995) Using diachronic linguistics in the classroom language In Modern

Language Journal Vol 111 Number 2 212-222

Rose Y and Demuth K (2006) Vowel epenthesis inloanword adaptation

representational ad phonetic considerations Lingua 116 1112-1139

Russel K (1995) Morphemes and candidates in Optimality Theory MS Rutgers

Optimality Archive

Salim J (2013) A contrastive study of English-Arabic noun morphology International

Journal of English linguistics 3 (3) 122-132

Sasa T (2009) Treatment of vowel harmony in Optimality Theory (Unpublished Phd

dissertation Iowa University)

Schutz J (2004) English loanwords in Fijian In Geraghty and Tent (2004b) 253-294

Scotton C and Okeju J (1972) Loanword integration in Ateso Anthropological

linguistics 14(9) 368-382

Selinker L (1972) Rediscovering inter-language New York Longman Group Ltd

Shabiro M (2015) A unique case of vowel harmony in English (Iambaste) Retrieved

from Michael Shapirorsquos wepsite ldquosemantic neostructuralism in peircean

linguisticsrdquo

Shademan S (2003) Epenthetic vowel harmony in Farsi (MA thesis University of

Califonia Los Angles)

Shilington K (1995) History of Africa Oxford Macmillan publishers limited

Shinohara S (2006) Perceptual effects In final cluster reduction patterns Lingua116

1046-1078

Shinohara S (1997) Analyse phonolgique de lrsquo adaptation Japonaisedes mots

entrangers (PhD dissertation Paris III)

Sing R (1981) On some redundant compounds in modern Hindi Lingua 56 p 345-351

Smolensky P (1996) The initial state and ldquoRichness of the Baserdquo in Optimality Theory

(Unpublished Ms Johns Hopkins University Baltimore (ROA-154))

Singabapha C (1998) The phonologization of English loanwords in Ikalanga and the

implications of this process for the English spelling and pronunciation by Ikalanga

speakers (BA dissertation University of Botswana)

Silverman D (1992) Multiple scansions in loanword phonology evidence fro

Cantonese Phonology 9289-328

Spencer A and Zwicky A (eds) (1998) The handbook of morphology Oxford

Blakwell publishers

Steel S (1995) Towards a theory of morphological information Language71 260-309

Taber C R (1979) French loanwords in Sango the motivation of lexical borrowing In

Hancock I (ed) Readings in creole studies Ghent scientific studies

Takagi N and Mann V (1994) A perceptual basis for the systematic phonological

correspondences between Japanese loanwords Journal of phonetics 22 243-356

Thomason S and Kaufman T (2008) Linguistic contact creolization and genetic

linguistics Berkeley University of California Press

Trubetzkoy N (1969) Principles of Phonology Berkeley and Los Angeles Califonian

University press

Uffmann C (2004) Vowel epenthesis in loanword phonology (PhD dissertation Philipps

University Morbug Germany)

Uffmann C (2013) Set the controls for the heart of the alternation Dahlrsquos Law in

Kitharaka Nordlyd 401323-337

Uffmann C (2001) Patterns of vowel epenthesis (not only) in Shona loanwords In

Anthony Dubach Green Caroline Feacutery amp Ruben van de Vijver (eds) Proceedings

of HILP 5 Universitaumlt Potsdam

Uffimann C (1993) Distinctive features an introduction (Lecture notes Stanford

Philipps University Marburg)

Vendelin I and Peperkamp S (2006) The influence of orthography on loanword

adaptations Lingua 116 996ndash1007

Verma K and Krishnaswamy N (1989) Modern linguistics an introduction Delhi

Weinreich H (1953) Language in Contact findings and problems The Hague Mouton

and Co Ltd

Werker J and Toes C (1984) Phonemic and phonetic factors in adult cross language

speech perception Journal of Acoustic Society of America 75 1866-1878

Whiteley W (1967) Loanwords in linguistic description a case study from Tanzania

East Africa Ranch I and Scott C (eds) Approaches in linguistic methodology

Mounton deGruyer

Whiteley W (1965) A practical introduction to Gusii Nairobi EALB

Witeley W (1960) Tense system of Gusii Kampala East African Institute of Research

Winter W (1992) Borrowing and non-borrowing in Walapi In EH Jahr (Ed)

Language contact theoretical and empirical studies Berlin Mounton deGruyer

Wu H (2006) Stress to tone a study of tone loans in Mandrine Chinese MIT working

Papers in linguistics 52 227-253

Wunderlich D (1996) Minimalist morphology the role of paradigms Yeaybook of

Morphology 1995 93-114

Yip M (2006) The symbiosis between perception and grammar in loanword phonology

Lingua116 950-975

Yip M (2002) Necessary but not sufficient Perceptual influences in loanword

phonology Journal of phonetic society of Japan 6 4-21

Yip M (1993) Cantonese loanword phonology and Optimality Theory Journal of East

Asian Linguistics 2 261-291

Zikmund W Babin J Carr J amp Griffin M (2012) Business research methods with

qualtrics printed access card Cengage Learning

Zivenga W (2009) Phonological and Morphological Nativization of English Loans in

Tonga (Unpublished PhD dissertation UNISA)

Zwicky A (1985) How to describe inflection In proceedings of BLSII ed by Mary

Niepokuj Mary vanClay Vassiliki Nikifondou and Deborah Feder 372-386

University of Califonia Berkeley

APPENDICES

Appendix I Interviewee profile form

1 Namehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

2 Agehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

3 Genderhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

4 Countyhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

5 Sub- Countyhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

6 Area of current residence (village)helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

7 Area of former residence (if any)helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

8 First languagehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

9 Knowlede of

(i) Kiswahilili languagehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

(ii) English languagehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

(iii) Any other language(s)

10 Occupationhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

11 Level of educationhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

12 School(s) attendedhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

Appendix II Interviewee consent form

I MrMrsMshelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip agree to

participate in the research exercise being conducted by Mr George Morara Anyona of Kisii

University I wish to state that he has made me aware of what he requires and have

voluntarily and willingly accepted to volunteer information pertaining to EkeGusii

language for purposes of the research

ID Nohelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

Sub-location helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

Countyhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

Datehelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

Signhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip

Appendix III Research instrument semi-structured interview guide

Interview questions in this study were based on the following thirteen (13) domains of life of the nouns

Domain of the nouns

Questions

1 FOOD AND NUTRITION

1 Ninki orantebie igoro yersquochindagera nersquobinyugwa mokoroisia nakoria aaiga amo na chinkaki mogochiooni

(What can you tell me about the food and drinks you prepare and sell here and the time you sell them)

2 HOUSE-HOLD APPLIANCES AND UTENSILS

1 Ntebo anko igoro yersquogasi yao yarsquokera rituko

(Narrate your daily work)

2 Ntebie ebinto bionsi bire nyomba aiiga gwansera ase mokorara mokorugera mogwesiberia mogoikaransa ( mention all the household things found in this house in the bedroom batheroom kitchen and sitting room)

3 AGRICULTURE 1 Aye norsquoyomo bwarsquobaremi aaiga koranche narrateigoro yorsquoboremi bwao (You are one of the farmers in this area Please tell me more about your farming activities)

2 Ntebi igoro ye chinchera chioboremi chigokorekana aaiga (Tell me more about the types of farming activities carried out in this area)

3 Ntebi ebimeria biria bigosimekwa aaiga Etugo rende (Tell me the type of crops that are grown in this area What about animals)

4TRANSPORT AND MOTORING

1 Ntebirsquoango igoro yersquogasi yao Ndi gwaansete korosia chigari (when did you become a mechanic)

2 Mechando ki okonyora (Have you ever faced any challenged)

(What challenges do you face)

3 Ntebirsquoanko emechando emenene yechigari okorosia

(Talk about the major mechanical problems you deal with )

4 Ntebianko buna ebioma aoao biegari bigokora emeremo

(Tell me how a vehicle parts work)

5 Ntebi aina ye chigari okorosi

(Tell me about the kinds of vehicles that come for repair)

5 HEALTH 1 Enkaki engana ngaki gwkorete agasi buna omonyagitari (For how long did you work as a nurse)

2 Ntebi buna enyagitari egokora egasi (Tell me about how a hospital generally works)

3 Ntebi igoro yersquogasi yaobuna omorwaria (Narrate to me about your work as nurse)

4 Ntebi igoro yarsquo marwaire acoria abanto aaiga (Tell me about themost prefernt diseases in this area)

5 Ntebirsquoank buna abarwaia aoao narsquobakoriersquogasi bersquonyagitari bagokora emeremo (Tell me about how the various hospital personel work)

6 RELIGION 1 Ntebirsquoanko ekanisa yao (Tell me the denomination you belong to)

2 Intebie igoro yersquo kanisa eyio nersquochinde omanyete) (Tell me more about this denomination and any other that you know of)

3 Teba ebinto biria bigokorekana ekninisa rituko roirsquogosasima (Narrate the activities that during your worshiping day)

7 EDUCATION 1 Ntebi igoro yorsquobogima bwao bworsquogosoma (Tell me about your education life)

2 Ntebi igoro yersquobinto birie bigokorekana esukuru rituko riersquosukuru (Narrate about the activities that take place in school in a normal day)

3 Teba igoro yarsquo baria bonsi bakobwaterana igoro yarsquomangana yersquosukuru (Talk about all the stakeholders in a school set up)

8 POLITICS 1 Kwabeire ime ya siasa amatuko amange Ntebirsquonko ebirogo biria abanasiasa bakorwanerera (You have been in politics for so long Tell me about the various political positions that politicians

vie for)

2 Nonyare kongeresa igoro ya siasa yersquonse (Can you tell me something about national politics)

9 LEGAL AFFAIRS

10ECURITY AND ADMINISTRATION

1 Gwakorire egasi yersquokoti ase enkaki enyinge Ntebi igoro yersquo chikoti chiarsquokenya buna chibangire na gokora egasi(You have worked in courts for long Tell me about the court system in Kenya)

2 Ntebi igoro yabakoriersquogasi bersquokoti (Tell me about the personnel of the courts)

1 Ntebi igoro yersquogasi yao (Tell me more aout your work)2 Ntebi buna obogambi bersquonse bobankire (Describe the structure of the national administration)

11INFORMATIONCOMMUNICATIONamp TECHNOLOGY

1 Egasi yao norsquokorosia ebito ebi Nkorosiorsquore binto binde otatiga ebio ndoche abuo Koranche ntebi igoro yersquobinto ebio( You earn aleaving by repairing thins things Do you repair anything else besides what I see on the shelves Please tell me more about them)2 Tell me about the changes you have witnessed over time concerning your work

12BUSINESS TRADE

13 CLOTHING

1 Koranche ntebi igoro yersquobiasara biao (Please tell more about your business) Ninki ogokora kera rituko as egasi eyio (What do on a daily basis)2 Ntebi igoro yersquobinto biria okogora na koonia aaiga (Tell me about the goods and services you buy and sell here)

1 Koraanche ntebi egasi yao (Please tell me what you do) Ntebi gochirsquome mono igoro yersquogasi eyio yao) Tell me more about your work2 Iyaankarsquoki egetaamba bunersquoke keraroisie (what type of clothes can such piece of clothe make) Naende gento kende(Anything else)

Appendix IV Raw data recordingtranscription form

Semantic domainhelliphelliphelliphellip1 (food and nutrition)

LOAN NOUN PHONOLOGICAL FORM MORPHOLOGICAL FORM SOURCE WORD FORM

Appendix V Raw data

Semantic domainhelliphelliphelliphellip1 (food drinks and nutrition) - By a Hotelier

LOAN WORD PHONOLOGICAL FORM MORPHOLOGICAL FORM SOURCE WORD FORM

burekibasitiranchisabaekarotichikarotiekabichichikabichiekekichikekiekerimuebirimuesotachisotagurukosieturunkiekokoripaipaiamapaipaiesigara

βurekiβasitiranchisaβaekarͻtitintikarͻtiekaβitintitintikaβitintiɛkɛkitintikɛkiekerimueβirimuɛsͻtatintisͻtaγurukosieturuŋkiekokoripaipaiamapaipaiesiγara

burek-i-basit-Iranch-isab-ae-karot-ichi-karot-ie-kabich-ichi-kabich-ie-kek-ichi-kek-ieke-rim-uebi-rim-ue-sot-achi-sot-agurukos-ie-turunk-ie-kok-ori-paip-aiama-paip-aie-sigar-a

breakfast brekfastlunch lʌndʒsupper sʌpǝcarrot kǝrǝtcarrots kǝrǝtscabbage kǝbɪdʒcabbages kǝbɪdʒizcake keɪkcakes keɪkscream krim

- -soda sɒdǝ

- -glucose glukǝʊzdrink drɪŋkcocoa kɒkǝʊpawpaw pǝʊpǝʊ-cigeratte sɪgǝret

Semantic domainhelliphelliphelliphellip2 (Household appliances and utensils) ndash By a house wife294

LOAN WORD PHONOLOGICAL FORM MORPHOLOGICAL FORM

SOURCE WORD FORM

etochichitochiekerasiebirasietamosichitamosiebirichichibirichioboranketiamaranketietaurochitauroerongrsquoIchirongrsquoiesatichisatiekebesitiebibesitiegotichigotiechaketichichaketiesikatichisikatiesogisichisogisieburaosichiburaosietankichitankiebetiruumuchibetiruumuekichenichikicheniebaturuumuchibaturuumuesinkichisinkieburasichiburasiegeita

ɛtͻtintitintitͻtintiekerasieβirasietamositintitamosiebiritintitintiβiritintioβoraŋketiamaraŋketietaurotintitauroɛrͻnŋitintirͻŋiesatitintisatiekeβesitieβiβesitieγotitintiγotietintaketitintitintaketiesikatitintisikatiɛsͻγisitintisͻγisieβuraositintiβuraosietaŋkitintitaŋkiɛβɛtiruumutintiβɛtiruumuekitintenitintikitintenieβaturuumutintiβaturuumuesiŋkitintisiŋkieβurasitintiβurasieγeita

e-toch-Ichi-toch-ieke-ras-iebi-ras-ie-tamos-ichi-tamos-ie-birich-ichi-birich-iobo-ranket-iama-ranket-ie-taur-ochi-tau-roe-rong-ichi-rong-rsquoie-sat-ichi-sat-ieke-besit-iebi-besit-ie-got-ichi-got-ie-chaket-ichi-chaket-ie-sikat-ichi-sikat-ie-sogis-ichi-sogis-ie-buraos-ichi-buraos-ie-tank-ichi-tank-ie-betiruum-uchi-betiruum-ue-kichen-ichi-kichen-ie-baturuum-uchi-baturuum-ue-sink-ichi-sink-ie-buras-ichi-buras-iegeita

torch tͻtinttorches tͻtintɪzglass glᴂsglasses glᴂsizthermos θǝmɒs- -fridge frɪdʒfridges frɪdʒɪzblanket blᴂŋkɪtblankets blᴂŋkɪtstowel tǝwɛltowels tǝwɛlzlong trouser lɒŋtrǝʊsǝlong trousers lɒŋtrǝʊsǝsshirt intɜtshirts intɜtsvest vɛstvests vestscoat kǝʊtcoats kǝʊts jacket dʒʌkɛt jackets dʒʌkɛtskirt skɜtskirts skɜtssocks sɒks

- -blouse blǝʊzblouses blǝʊzɪztank tᴂŋktanks tᴂŋks bedroom bɛdrum bedrooms bɛdrumskitchen kɪtintǝn kitchens kɪtintǝns bathroom bᴂethrumbathroom bᴂethrumzsink sɪŋksinks sɪŋksbrush brʌintbrushes brʌintɪzgate geɪt

ebiitaeswentachiswentaesitingrsquoiruumuchisitingrsquoIruumuekabatichikabatiebiichachibiichaegasichigasiesobachisobaesitochisitoebesenichibeseniebatiraepogisiekotoni

eβiitaeswentatintiswentaesitingrsquoiruumutintisitiŋiruumuekaβatitintikaβatieβiichatintiβiitintaeγasitintiγasiesoβatintisoβaesitotintisitoɛβɛsɛnitintiβɛsɛnieatiraepͻγisiɛkͻtoni

e-biit-aeswentachi-swent-ae-sitingrsquoiruum-uchi-sitingrsquoiruum-ue-kabat-ichi-kabat-ie-biich-achi-biich-ae-gas-ichi-gas-ie-sob-achi-sob-ae-sito-ochi-sito-oe-besen-ichi-besen-ie-atir-ae-pogis-ie-koton-i

gates geɪtssweater swetǝsweaters swetǝzsitting room sɪtɪŋrum sitting rooms sɪtɪŋrumscupboard kʌbǝd cupboards kǝbǝdzpicture pɪktintǝpictures pɪktintǝzgas gᴂzgases gᴂzɪzsofa sǝʊfǝsofas sǝʊfǝzstore stͻstores stͻsbasin beɪsnbasins beɪsnsbottle bɒtlbox bɒkscotton kɒtn

Semantic domainhelliphelliphelliphellip3 (Agriculture farming)- By an agricultural officer

SOURCE WORD FORM

ekeragitaebiragitaerainichirainiebambuchibambuomokirigachaabakirigachaegurubarochigurubaroeekachiekaegiratichiegiretiesirasichisirasiegwayachiwayaetiibuchitiibuebutichibuti

ekeraγitaeβiraγitaeraini tintirainieβambu tintibambuomokiriγatintaaβakirigatintaegurubarotintiγuruβaroeeka tintiekaeγiretitintiegiretiesirasi tintisirasi|eγuaya| eγwaya|tintiγuaya| tintiwayaetiiβutintitiiβueβutitintiβuti

eke-ragit-aebi-ragit-ae-rain-ichi-rain-ie-bamb-uchi-bamb-uomo-kirigach-aaba-kirigach-ae-gurubar-ochi-guruba-roe-ek-achi-ek-ae-giret-ichi-e-giret-ie-siras-ichi-siras-ie-gway-achi-gway-ae-tiib-uchi-tiib-ue-but-ichi-but-i

tractor trʌktǝtractors trʌktǝzline lainlines lainspump pʌmppumps pʌmpsagriculture officeragriculture officerswheel barrow wilbǝrǝʊswheelbarrows wilbǝrǝʊacre eɪkǝacres ekǝs grade greɪdgrade onesslush slʌintslushes slʌintizwire wǝɪǝwires wǝɪǝsdip dɪpdips dɪpsfeet fitfeets fits

Semantic domainhelliphelliphelliphellip4 (transport and motoring) ndash By a motor mechanic

erorichiroriesitarinkichisitarinkieburekichiburekiekerachiebirachiekaachikaaebetiroriomonterebaabanterebaomomakanikaabamakanikaeboritichiboritiebetirichibetiriechekichichekietagisichitagisiomokondagitaabakondagitaetiseriemasinichimasiniegerechichigerechiomotokaemetokaetaerichitaeri

eroritintiroriesitarinkitintisitarinkieβureki tintiβurekiekera tintieβiratintiekaatintikaaeβetiroriͻmͻntɛrɛβaaβantɛrɛβaomomakanikaaβamakanikaɛβͻrititintiβͻritiɛβɛtiritintiβɛtirietintɛkitintitintekietaγisitintitaγisiͻmͻkͻndaγitaaβakͻndaγitaetiseriemasinitintimasiniegaratintitintigaratintiͻmͻtͻkaɛmɛtͻkaetaeritintitaeri

e-ror-Ichi-ror-ie-sitarink-ichi-sitarink-ie-burek-ichi-burek-ieke-rach-iebi-rach-ie-ka-achi-ka-ae-betiror-iomo-ntereb-aaba-ntereb-aomo-makanik-aaba-makanik-ae-borit-ichi-borit-ie-betir-ichi-betir-ie-chek-ichi-chek-ie-tagis-ichi-tagis-iomo-kondagit-aaβa-kondagit-ae-tiser-ie-masin-ichi-masin-ie-garech-ichi-garech-iomo-tok-aeme-tok-ae-taer-ichi-taer-i

lorry lɒrilorries lɒrizsteering stɪǝrŋ-break breɪkbreaks breɪksclutch klʌtintclutches klʌtintɪzcar kɑcars kɑpetrol petrǝldriver draɪvǝdrivers draɪvǝzmechanic mǝkaelignɪkmechanics mǝkaelignɪksbolt bɒlt bolts bɒltsbattery baeligtrɪbatteriesbaeligtrɪzjerk dʒɜk jerks dʒɜks taxi taeligkstaxis taeligksɪzconductor kǝndʌktǝconductors kǝndʌktǝzdiesel dizlmachine mǝintinmachines mǝintinsgarage gaeligrɑʒgarages gaeligrɑʒɪzmotor car mǝʊtǝ kɑ motor cars mǝʊtǝ kɑztile taɪltiles taɪls

Semantic domainhelliphelliphelliphellip5 (Health) ndash By Health practitioner- nurse

mareritibiiekerinikiebirinikimatenetichimatenetiomonasiabanasiemocherichimocherieambiurensitintiambiurensieteresachiteresaomoteresaabateresaebandechi

marɛritiβiiekerinikieβirinikimatenetitintimatɛnɛtiomonasiaβanasiemͻtinteritintimͻtintɛri |eambiurenzi|eambjurenzi|tintiambiurenzi|tintiambjurenziɛtɛrɛsatintitɛrɛsaͻmͻtɛrɛsaaβatɛrɛsaeβandetinti

marer-Itib-iieke-rinik-iebi-rinik-ie-matenet-ichi-matenet-iomo-nas-iaba-nas-iemo-cher-ichi-mocher-ie-ambiurens-ichi-ambirens-ie-teres-achi-teres-aomo-teresaaba-teres-ae-bandech-i

malaria mǝleǝriǝtp tipiclinic klɪnɪkclinics klɪnɪksmaternity mǝtɜnǝti maternities mǝtɜnǝtiznurse nɜsnurses nɜsɪzmortuary mͻtintǝrimortuaries mͻtintǝrizambulanceaeligmbjǝlǝnsambulances aeligmbjǝlǝnsɪzdresserdressers dresser dresǝdressers dresǝz bandage baeligndɪdʒ

Semantic domainhelliphelliphelliphellip6 (religion) - By Church elder

SOURCE WORD FORM

biechiikatorikiesitieeomobataababataomobasitaababasitaebukuchibukuekibotichikibotiegiitaebiitaekerisimasiebirisimasirisakaramentiamasakaramentiomokatorikiabakatorikikirisitoekorasichikorasiekwayachikwayaomokiristoabakirisitoomoesitieeabaesitieeomobisopuababisobu

βietintiikatorikiɛsitieeomoβataaβaβataomoβasitaaβaβasitaeβukutintibukuekiiβͻtitintikiiβͻtieγiitaeβiitaekerisimasieβirisimasirisakaramentiamasakaramentiomokatorikiaβakatorikikirisitoɛkͻrasitintikͻrasi|ekuaya|| ekwaja|tintikuaja| tintikwajaomokiristoaβakirisitoͻmͻɛsitieeaβaesitieeomoβisͻpuaaisͻpu

biech-iikatorik-iesit-ieeomo-bat-aaba-bat-aomo-basit-aaba-basit-ae-buk-uchi-buk-ue-kibot-ichi-kibot-ie-giit-ae-biit-aeke-risimas-iebi-risimas-iri-sakarament-iama-sakarament-iomo-katorik-iaba-katorik-ikirisit-oe-koras-ichi-koras-ie-kway-achi-kway-aomo-kirist-oab-akirisit-oomo-esit-ieeaba-esit-ieeomo-bisop-uaba- bisop-u

PAG pieɪdʒicatholic kaeligθɒlɪkSDA esdieɪfather fǝethǝfathers fǝethǝzpastor pʌstǝpastors pʌstǝzbook bʊkbooks bʊkskey board kibɒdzkey board kibɒdzguitar gɪtɑguitars gɪtɑzChristmas krɪmǝs-sacrament saeligkrǝmǝntsacraments saeligkrǝmǝntsa catholic aeligkaeligθlɪkcatholics aeligkaeligθlɪksChrist kraɪstchorous kɒrǝzchorouses kɒrǝzɪzchoir kwaɪǝchoirs kwaɪǝza Christian aelig krɪstintǝnChristians krɪstintǝnsan SDA aelign SDAs es dieɪzbishop bɪintǝpbishops bɪintǝps

Semantic domainhelliphelliphelliphellip7 (education) - By an educationst

SOURCE WORD FORM

esukuru esukuru e-sukur-u school skul300

chisukuruekerasiebirasieburakibotichiburakibotichookachichookaeyunibomuchiyunibomuenasarichinasarieburemarichiburemariesekondarichisekondarieyunibasitychiyunibasitiekorechichikorechiebaerichibaerietasitachitasitaebenchichibenchietamuchitamuetigirichitigiriekosichikosiewikichiwiki

tintisukuruekerasieβirasieβurakiβͻti tintiβurakiβͻtiɛtintͻͻka tintitintͻͻkaeyunibomu tintiyuniβͻmuenasari tintinasarieβurɛmari tintiβurɛmariɛsɛkͻndari tintisɛkͻndarieyuniβasiti tintiyuniβasitiekoretinti tintikoretintieβaeri tintiβaerietasita tintitasitaeβendƷitintiβendƷietamutintitamuetiγiritintitiγiriekosi tintikosi|eγuiki|eγwiki|tintiγuiki| tintiwiki

chi-sukur-ueke-ras-iebi-ras-ie-burakibot-ichi-burakibot-ie-chook-achi-chook-ae-yunibom-uchi-yunibom-ue-nasar-ich-inasar-ie-buremar-ichi-buremar-ie-sekondar-ichi-sekondar-ie-yunibasit-ichi-yunibasit-ie-korech-ichi-korech-ie-baer-ichi-baer-ie-tasit-achi-tasit-ae-bench-ichi-bench-ie-tam-uchi-ta-uemdashtigir-ichi-tigir-ie-kos-ichi-kos-ie-wik-ichi-wik-i

schools skulzclass klaeligsclasses klaeligsɪzblackboard blaeligkbͻdblackboards blaeligkbͻdzchalk tintͻkpieces of chalk tintͻkuniform junfͻmuniforms junfͻmznursery nɜsnurseries nɜsɪzprimary praɪmǝrɪprimaries praɪmǝrɪzsecondary sekǝndri-university junɪvǝsɪtɪuniversities junɪvǝsɪtɪscollege kɒlɪdʒcolleges kɒlɪdʒɪzfile faɪlfiles faɪlsduster dʌstǝdusters dʌstǝzbench bendʒbenches benʒɪzterm tǝmterms tǝmzdegree dɪgridegrees dɪgrizcourse cɒscourses cɒsizweek wikweeks wiks

Semantic domainhelliphelliphelliphellip8 (politics governance and security) ndash By an assistant chief

SOURCE WORD FORM

eburesitentichiburesitentiomoburesitentababuresitentiegabanachigabanaomogabanaabagabanaesenetachisenetaomosenetaabasenetaekansarachikansaraomokansaraabakansaraetigitetachitigiteteekambichikambiesitesenichisitesenietibisonichitibisonierokesenichirokesenietisiturigitichitisiturigitiomoepiabaepiomoporisiabaporisiekomitiichikomitiierekotichirekotiomotisiiabatisiiomotiooabatiooomochibuabachibuomosabuchibuabasabuchibuomokirakiabakiraki

eβuresitenti tintiβuresitentiomoβurɛsitɛntiababurɛsitɛnti

omoγaβanaaβaγaβana

omoseenetaaβaseeneta

omokanzaraaβakanzaraetiγiteta tintitiγitetaekambi tintikambiesiteseni tintisitesenietiβisoni tintitiβisonierokeseni tintirokesenietisituriγiti tintitisituriγitiomoepiaβaepiomoporisiabapͻrisiɛkͻmitii tintikͻmitiiɛrɛkͻtitintirɛkͻtiomotisiiaβatisiiomotiooaβatiooomo tintiβuaβac tintiβuomosaβuchiβuaβasaβutintiβuomokirakiaβakiraki

e-buresitent-Ichi-buresitent-iomo-buresitent-iaba-buresiten-ti

omo-gaban-aaba-gaban-a

omo-senet-aaba-senet-a

omo-kansar-aaba-kansar-ae-tigitet-achi-tigitet-ae-kamb-ichi-kamb-ie-sitesen-ichi-sitesen-ie-tibison-ichi-tibison-ie-rokesen-ichi-rokesen-ie-tisiturigit-ichi-tisiturigit-iomo-eep-iaba-eep-iomo-poris-iaba-poris-ie-komit-iichi-komit-iie-rekot-ichi-rekot-iomo-tis-iaba-tis-iomo-tio-oaba-tio-oomo-chib-uaba-chib-uomo-sab-u-chib-uaba-sab-u-chib-uomo-kirak-iaba-kirak-i

president prezɪdǝntpresidents prezɪdǝntspresidentprezɪdǝntpresidents prezɪdǝnts

governor gʌvǝnǝgovernors gʌvǝnǝz

senator sɪnaeligtǝsenators sɪnaeligtǝz

counsillor kaʊnsǝlǝcounsillors kaʊnsǝlǝzdictator dɪkteɪtǝdictators dɪkteɪtǝzcump kʌmpcumps kʌmpsstation steɪintnstationssteɪintnsdivision dɪvɪintndivisions dɪvɪintnslocation lɒkeɪintnlocations lɒkeɪintnsdistrict dɪstrɪkdistricts dɪstrɪksan AP eɪpiAps eɪpisa police(manwoman) police(manwoman)committee kɒmiticommitteeskɒmitisrecord rekͻdrecords rekͻdza DC diziDCs dziza DO diǝʊDos diǝʊza chief tintifchiefs tintifssub chief sʌbtintifsub chiefs sʌbtintifsclerk klǝkclerks klǝks

erumandechirumandesekiuritiomoturetiabaturetirisabuamasbu

erumande tintirumandesekiuritiomoturetiaβaturetirisauaamasau

e-rumand-e chi-rumand-esekiuritiomo-turet-i aba-turet-i ri-sabua-masabu

remand rɪmǝndremands rɪmǝndssecurity sɪkjʊǝrǝtitraitor treɪtǝtraitors treɪtǝzreserve rɪsɜvreserves rɪsɜvz

Semantic domainhelliphelliphelliphelliphelliphellip9 (legal affairs) ndashBy court officer

SOURCE WORD FORM

echachichichachiomochachiabachachiemachisituretichimachisituretiomomachisituretiabamachisituretiekotichikotiomoroyaabaroyaomopurosekiutaabapurosekiutaebainichibainiekesichikesi

etintatintitintitintatintiomotintatintiaβachatintiematintisituretitintimatintisituretiomomatintisituretiaβamatintisituretiekoti tintikotiomorojaaβarojaomopurosekiutaaβapurosekiutaeβaini tintiβainiekesi tintikesi

e-chachichi-chachiomo-chach-Iaba-chach-ie-machisituret-ichi-machisituret-iomo-machisituret-iaba-machisituret-ie-kot-ichi-kot-iomo-roy-aaba-roy-aomo-purosekiut-aaba-purosekiut-ae-bain-ichi-bain-e-kes-ichi-kes-i

judge dʒʌdʒjudge dʒʌdʒɪzjudge dʒʌdʒjudges dʒʌdʒɪzmargistrate maeligdʒɪstreɪtmargistrates maeligdʒɪstreɪtsmargistrate maeligdʒɪstreɪtmargistrates maeligdʒɪstreɪtscourt kͻtcourts kͻtslawyer lͻjǝlawyers lͻjǝzprosecutor prɒsɪkjutǝprosecutors prɒsɪkjutǝzfine faɪnfines faɪnzcase keɪscases keɪsɪz

Semantic domainhellip10 (information communication amp technology) ndash By an IT expert

SOURCE WORD FORM

emesechichimesechieretiochiretioetibiichitibiiemobaerichimobaerieirioochiiriooekombiutachikombiutaesenemachisenema

e-mesech-Ich-imesech-ie-ret-iochi-ret-ioe-tib-iichi-tib-iie-mobaer-ich-imobaer-ie-irio-ochi-irio-oe-kombiut-achi-kombiut-ae-senem-achi-senem-a

message meseɪdʒmessages meseɪdʒizradio reɪdɪǝʊradios reɪdɪǝʊsTV tiviTVs tivizmobile mǝbaɪlmobiles mǝbaɪlzaerial earɪǝlaerials earɪǝlscomputer kɒmjutǝcomputers kɒmjutǝzcinema sɪnǝmǝcinemas sɪnǝmǝz

Semantic domainhelliphelliphelliphellip11 (businesstrade) By a Business man

SOURCE WORD FORM

ebucheerichibucheeriekirochikirorisitiamarisitieoterichioteriepaachipaaepiachipiaetaonichitaoniesubamaketichisubamaketietureichitureieresesichiresesiekiretiebiretietasanichitasaniesimitichisimitiechenchichichenchirinotiamanotiesirinkichisirinkiemarigitichimarigitiebankichibankiecheki

ebutinteeritintibutinteeriekiro tintikirorisitiamarisitieoteri| tintioteri| tintjoteriepaa tintipaa|epia| epja tintipiaetaoni tintitaoniesuβamaketi tintisuβamaketieturei tintitureieresesi tintiresesiekeretieβiretietasani tintitasaniesimiti tintisimitietintendƷi tintichendƷirinotiamanotiesirinki tintisirinkiemariγiti tintimariγitieβaaŋki tintiβaaŋkiɛtintɛkitintitintɛki

e-bucheer-Ichi-bucheer-ie-kir-ochi-kir-ori-sit-iama-risit-ie-oter-ichi-oter-ie-pa-achi-pa-ae-p-iachi-pi-ae-taon-ich-itaon-ie-subamaket-ichi-subamaket-ie-ture-ichi-ture-ie-reses-ichi-reses-ie-kiret-ie-biret-ie-tasan-ichi-tasan-ie-simit-ichi-simit-ie-chench-ichi-chench-iri-not-iama-not-ie-sirink-ichi-sirink-ie-marigit-ichi-marigit-ie-bank-ichi-bank-ie-chek-ichi-chek-i

butchery bʊtintǝrɪbutcheries bʊtintǝrɪzkilo kɪlǝʊkilos kɪlǝʊzreceipt rɪsitreceitsrɪsitshotel hǝʊtelhotels hǝʊtelsbar bɑbars bɑzbeer bɪǝ-town taʊntowns taʊnzsupermarket supǝmǝkeɪtsupermarkets supǝmǝkeɪtstray teɪtrays teɪz license lɪasǝnslicenses lɪasǝnsɪzcrate kreɪtcrates kreɪtsdozen dɒzndozens dɒznzcement sɪmǝntcementssɪmǝntschange tinteɪndʒ-a note nǝʊtnotes nǝʊtsshilling intɪlɪŋshillings intɪlɪŋzmarket m ɑkɪtmarkets m ɑkɪtsbank baeligŋkbanks bank baeligŋkscheque tintekcheques tinteks

Semantic domainhelliphelliphelliphellip13 (sports) ndash By a youths officer

SOURCE WORD FORM

gemusietiririchitiririegorichigorienetichinetiesibotichisibotieribariichiribariiemeratonichimaratonigwokingrsquoi resiekabutenichikabuteniomokabuteniabakabuteni

γemusietiriri tintitiririegori tintigoriɛnɛti tintinetiesβboti tintisiβotieriβarii tintiriβariiemeratonitintimaratoni|γuͻkiŋiresi| gwͻkiŋresiekaβutenitintikaβuteniomokaβuteniaβakaβuteni

gemus-Ie-tirir-ichi-tirir-ie-gor-ichi-gor-ie-net-ichi-net-ie-sibot-ichi-sibot-ie-ribar-iichi-ribar-iieme-raton-ichi-maraton-igwoking-i- res-ie-kabuten-ichi-kabuten-iomo-kabuten-iaba-kabuten-i

games geɪmzdrill drɪldrillsdrɪlsgoal gǝʊlgoals gǝʊlznet nǝtnets nǝtssport spͻtsports spͻtsfile faɪlfiles faɪlsmarathon maeligrǝӨǝnmarathonsmaeligrǝӨǝnzwalking race wͻkiŋreɪscaptain kaeligpteɪncaptainkaeligpteɪnscaptain kaeligpteɪncaptainkaeligpteɪns

Appendix VI Illustrations of phonological nativization

Source word form Nativized form phonological proces

christmas krɪsmǝs [ekirisimasi] segment change

taxi taeligksɪ [etagisi] segment change

sofa sǝʊfǝ [esoba] monophthongization

wire wǝɪǝ [ewaya] monophthongizaiton

finen faɪn [eaini] bilabialization

vest vest [eesiti] bilabialization

store stͻ [esitoo] expenthesis of [i]

fine faɪn [ebaini] epenthesis of [i]

coat kǝʊt [eγoti] vowel harmonization

basin baeligsn [εεsεni] vowel harmonization

Pastor pʌstǝ [asita] stop fricativization

bolt bͻlt [ͻriti] stop fricativization

drink drɪŋk |eturun-γi|rarr[eturuŋgi] fricative defricativization

camp kaeligmp |ekan-i|rarr[ekembi] fricative defricativization

bank baeligŋk [eeŋgi] voice dissimilation

location lǝʊkeɪintn [rokeHseni] stress tonemization

degree dɪgri [tigiHrii] stress tonemization

school skul [sukuru] syllable change

Appendix VII Illustrations of morphological nativization

Source word form Nativized form Morphological process

scout o-mo-sikaoti nominal classification (1)

scout-s a-ba-sikaoti nominal classification (2)

motor-car o-mo-tokaa nominal classification (3)

motor-car-s e-me-tokaa nominal classification (4)

torch (very big) ri-toochi nominal classification (5)

torch-es ama-toochi nominal classification (6)

school (very small) e-ke-sukuru nominal classification (7)

(small school (deminution))

school-s e-bi-sukuru nominal classification (8)

(small schools (deminition))

record ε-rεkͻti nominal classification (9)

record-s chi-rεkͻti nominal classification (10)

room (very small) aka-ruumu nominal classification (12)

deminution

room-s (very small) obo-ruumu nominal classification (14)

ticket e-tiketi prefixication

ticket-s chi-tiketi prefixication

governor o-mo-gabana pre-prefixication

governor-s a-ba-gabana pre-prefixication

Appendix VIII Enchoro sub-location villages

1 Sarama

2 Nyagenke

3 Ikaraancha

4 Esuguta

5 Nyakoria

6 Nyankarankania

7 Nyantaro

8 Nyagaachi

9 Motagara

10 Mariba A

11 Mariba B

12 Chumura

13 Mosobeti

14 Getukora

15 Enchoro

16 Ikarancha

17 Nyando

Appendix IX Research Clearance Permit

Appendix X Research Authorization Letter

Appendix XI Memo from the Registrar Research and Extension Kisii University toNACOSTI

Appendix XII Plagiarism Report

07-Nov-2017 10 35AM (UTC + 0300)

Appendix XIII Publication

DECLARATION

COPYRIGHT

ABSTRACT

DEDICATION

ACKNOWLEDGEMENT

TABLE OF CONTENTS

ABBREVIATIONS

SYMBOLS

CHAPTER ONE

INTRODUCTION

The sub-section begins with a brief description of the languages under focus in which an outline of the languages in terms of where they are spoken by who and the language families to which they belong Further it is shown how and why the two languages came into contact among other highlights This is then followed the background of the study

CHAPTER TWO

20 Introduction

CHAPTER THREE

30 Introduction

31 Research design

32 Research site

36 Data analysis

CHAPTER FOUR

40 Introduction

EkeGusii syllable structure is comparable to that of English While EkeGusii is a (V)CV language as shown in (31) the syllable typologies given in table (3) above show that English is a (C) (C) V (C) (C) syllable type of language Thus Onsets Codas Onset clusters and Coda clusters are all optional in English They may or may not occur depending on the nature of the word as shown in (33)

41412 Contour tones in EkeGusii

While EkeGusii is a tone language as has been discussed above English is a stress language Verma and Krishnaswamy (1989) define stress as the intensity or prominence with which a syllable is produced Thus in the production of a word or any syntactic structure (sentence or phrase) there is always a syllable which is produced with prominence It is this prominence that phonologists have come to refer to as stress

41522 The Prefix

42123 Nativization of English l

4311 The role of semantics in morphological nativization of English nouns in English

4315 Nativization in classes 5 eri-rii- 6 ama- 7 eke- 8 ebi- 12 aka- and 14 obo-

CHAPTER FIVE

51 Summary

52 Conclusions

53 Recommendations

REFERENCES

APPENDICES

Appendix V Raw data

DECLARATION

DECLARATION BY THE CANDIDATE

This thesis is my original work and it has not been submitted in this or any other university

known to me

Signhelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip Datehelliphelliphelliphelliphelliphelliphelliphelliphellip

George Morara Anyona

DAS13600102014

DECLARATION BY THE SUPERVISORS

This thesis has been submitted with our approval as university supervisors

Dr David O Ongarora PhD

Department of Linguistics Maseno University

Dr Evans G Mecha PhD

Department of Lit Lang amp Ling Kisii University

dishonesty

APPROACH

degree

(Candidate)

(Supervisor 1)

(Supervisor 2)

COPYRIGHT

ABSTRACT

DEDICATION

ACKNOWLEDGEMENT

deep indebtness

institutions

TABLE OF CONTENTS

DECLARATIONii

COPYRIGHTv

ABSTRACTvi

DEDICATIONvii

ACKNOWLEDGEMENTviii

TABLE OF CONTENTSx

CHAPTER ONE

INTRODUCTION1

CHAPTER TWO

20 Introduction12

CHAPTER THREE

30 Introduction40

32 Research site41

36 Data analysis43

CHAPTER FOUR

40 Introduction46

41522 The Prefix125

obo-265

CHAPTER FIVE

51 Summary276

52 Conclusions281

REFERENCES285

APPENDICES307

CHARTS

ABBREVIATIONS

AFX affix

APP applicative

C consonant

EVAL evaluator

FV final vowel

GEN generator

IND indicative mood

NC nasal Consonant

OM object Marker

P nominal Person

PL plural

SG singular

SM subject Marker

RT root of a word

TNS tense

V vowel

HTS high tone spans

correspondent tone

SYMBOLS

σ Syllable node

Optimal candidate

Fatal violation

Phonology

TABLES

TABLEAUX

CHARTS Page

FIGURES Page

xxviii

CHAPTER ONE

INTRODUCTION

spoken in Kenya

language

EkeGusii undergo

English nouns

into EkeGusii

this study

recipient language

CHAPTER TWO

20 Introduction

this study

phonology

is concerned

(Keith1997105)

EkeGusii

language

characteristics

Input GEN

Source Kager (1999)

chellip

gtgt gtgt

Output

correspondent tone

in tableau (1)

a) intip

b) skip

generative theories

are novel

constraints

different languages

CHAPTER THREE

30 Introduction

31 Research design

32 Research site

as follows

where it occurs

36 Data analysis

INPUT GEN OUTPUT

CHAPTER FOUR

40 Introduction

Time (s)0 0136

quency (

5000 ε

Time (s)0 007152

5000sound ee

Time (s)0 03045

Time (s)0 008388

quency (

5000sound ou

Time (s)0 007639

5000sounnd i

Time (s)0 1189

0594285714u

Time (s)0 007639

5000sounnd ii

Time (s)0 0136

ency (

5000ͻ

2500 2000 1500 1000 500

having a low F2)

4) English vowels

a) pure vowels

i heed feel bead

ɪ hid fill bid

e head fell bed

ᴂ had bad mad

ɑ hard bard par

ɒ hod bod

ʊ hood full

ʌ but cut hut

ǝ accept father

b) Dipthongs

eɪ fail bayed pay

ͻɪ foil boy

ǝʊ hoe foal

ɪǝ beard beer

ʊǝ poor

c) Triphthongs

ǝʊǝ mower slower

OrsquoConnor 1967)

Vowel Description

ᴂ ε

ͻ ͻͻ

languages

Table (2) above

Vowel F1 F2

i 540 2450e 730 2250ͻ 750 1250u 520 1100

Vowel F1 F2

i 280 2250e 400 1920ͻ 590 850u 310 890

DIPHTHONG

centring closing

(Ongarora 1996)

b) V1 V2 V3

c) V1 V2 V3

+HIGH])

Continuant β s γ

Obstruent [b] t k

Affricate c

Nasal m n ntilde ɳ

Glide w y

[Cw] [Cy]

Continuant ɸ s γ

(tril) r

Obstruent [b] t k

[d] [g]

Affricate [dʒ]

Nasal m n ɳ ŋ

Approximant [w] j

these processes

below demonstrates

[m b w a t e]

Output [wata]

Output sjeka

McCarthy (20079)

consonant glide [j]

c vcv c cvcv

clusters

a) rwana σ σ σ

C V V C V

rw u a n a

b) rwana σ σ

cc v c v

rw a n a

sought to answer

enda e nda

V NC V

e nd a

CVV CV C CVVC

node as figure (11)

σ σ σ σ

m o n t o m o n t o

(200290)

n-to- taatint -ɛ

F-1p-fetch-FV

[ntotaatinte]

Fndash 2p1-fetch-fv

[mmotaatinte]

c) in-a-taatint-e

F-3p-fetch-fv

[mbataatinte]

Suffixation -ee

Prefixation n-e-

figure (12)

[word [+Nasal]

Suffixation -γore

Prefixation n-γore

Output [ŋgore]

N C V C V

ŋ g o r e

CC V C V

ng o r e

syllables

i) ink rarr iŋk

cluster

X (C)CV(C) Dakota

XO CV(C)(C) Klamath

X CV(C) TemierR X

O _ (C)VC Arabela

X CV Senufo

X (C)V(C) Turkish

C V C V

γ e t ii

i) omote

vc vcv σ σ σ

v c v c v

o m o t e ii) a aa

vv σ σ

in this position

respectively

vowel word

i e letter ltegt

ǝ a ɑ are

ͻ ɜ err

eɪ a letter ltagt

aɪ i letter ltigt

ǝʊ owe

ɪǝ ear

eǝ air

others do not

Cruttenden (2011)

m+ j mew

n+ j new

l+ j lure leau

v+j view

h + j hew

C V V C V

γ a k a -γaaka

C V C V

γ a k a -γaka

L L H L

k כ bεlε mbu mba

H H HH

observation

H L H L H H L HLH

tone (HL)

shown in (38)

wik weak kɒntrʌkt

ɒdǝlɪ orderly

Singular Plural

meaning

(a) (b)

i) omot e σ

vc vc v σ σ σ

v c v c v

o m o t e

- mo- nto -βa- nto

-tata -βa- tata

rule as in (7)

vowel e

41522 The Prefix

previous section

aug- 3psg- root

aug-3pl- root

aug- 3psg- root

aug- 3pl- root

(i) Class 3

form form form

ii) Class 4

Class 9 e-

Class 10 chin-

Class 9a en-

Class 10a chin-

in (52)

Singular plural

(a) 7 eke - 8 eβi-

12 akas - 8 eβi-

12 aka- 14 oβo-

(b) 5 eri- 6 ama-

a) Diminution

a- ka- mo- onto

aug- 12- 1 - person

[akamoonto]

[akamoonto aγake]

[ekemoonto eγeke]

c) abusive

d) pejorative

phonology

(2001)

a [krɪsmǝs]

b [kirisimasi]

a [krɪsmǝs]

b[kirisimasi]

fatal violation

rounded low]

a [intɜt]

b esati

a [intɜt]

b [esati]

a [glɑs]

b [ekerasi]

a [glɑs]

b [ekerasi]

produce

a [sɪneɪt]

b [eseneti]

a [sɪneɪt]

b [eseneti]

ǝ [a]

a [lǝʊkeɪintn]

b [erokeseni]

a [lǝʊkeɪintn]

b [erokeseni]

English [aɪ]

a [waɪǝ]

b [waja]

a [waɪǝ]

b [waja]

a) frɪdʒ

b) eβuritinti

a) frɪdʒ

b) eβuritinti

a) skafu

b) sikava

minimal if at all

lsquothermosrsquo

a) Өɜməs

b) tamosi

c) ethɜməs

a) Өamosi

b) tamosi

c) ethɜməs

[filimu]

a) claeligs

b) ekerasi

c) ekelasi

a) sukuru

b) sukulu

in (85)

a) zirəʊ

b) esiro

a) esiro

b) zirəʊ

a) intɜt

b) esati

a) intɜt

b) esati

a) baeligg

b) epaγi

a) baeligg

b) epaγi

O N O N

C V C V

m a m a

N O N O N

V C V C V

o m o t e

CC V CC

s t eɪ nz

Input mama

a) mama

b) a ama

c) maa

a) ǝʊ

b) tǝʊ

c) ǝʊǝ

a) teɪ

b) teɪk

c) teɪki

a) seɪn

b) siteɪnsɪ

c) steɪns

shown in (90)

English EkeGusii

σ σ σ

O N O N O N

CC V C V C V

s t ͻ s i t o

nativized form

a) sukuru

a) skuru

c) sukuru

EkeGusii

English EkeGusii

σ σ σ σ

O N C O N O N O N

CC V C C V C V C V

Samoan Sranan

English EkeGusii

σ σ σ

O V C O N O N

C N C C V C V

s u t s u t i 192

a) tͻtint

b) tͻtinti

a) tͻtint

b) tͻͻtinti

Input t-n-ak-ol

a) t-n-ak-ol

b) t i nagͻl

a) noŋkomati

b) no-n-koma-i

illustrated by (96)

L H L H

L L H L H

L L H L

L H L H

c) Dholuo

Input blaeligŋkɪt

Input oo-raŋget-i

FLOAT MAX-IO

IDENT-IO

DEP-IO SPEC

a) okokoro

b) okoγoro

(transfer)

a) baeligŋk

b) eŋgi

c) e - ki - ar - a d) e - bi - ar- a

the plural form

11 boy

finger

ii) |r| rarr [d]

shown in (107)

Output [kambi]

a kambi

b kani

c kanbi

defricativization

are both bilabials)

a) eteγisi

b) tᴂksi

[γ] as the output

sub-section

in tableau (42)

a ɸaranda

b vǝrᴂndǝ

i) drink

(ii) veranda

Input e-+βaranr-a

English EkeGusii

veranda baranda

tableau (43)

a ɸaranda (unit)

UF SF Gloss

a ɛntɛrɛa

b draɪvǝ

follows

a nkooko

b e-nkooko

a e-nkooko

b nk-ooko

b boy-s

a boy-s

Input blanket-s

(50) as follows

a blanket

b ama- ranketi

c blanket-s

nominal class

form as in (118)

objects

below show

Input tractor-s

analysis

L RT R)) STRPRES

a prefix

Input o- mo- te

aug 33PSG tree

output [omote]

b record-s

c s-record

at syllable margins

in tableau (55)

Input recordrekͻ d

a rekͻd-s

b s-rekͻd

c rekͻd

933PSG- record

a e-rekoti

b record

c record-s

highly ranked

103PPL record

a chi-rekoti

b record

c record-chi

violations

are 12 34 14 6 7 8 and 12

a o-mo- chiibu

b chief

c chief-s

Input chief

a o-mo- chiibu

b chief

c chief-s

prefix

denote plurals

classes 3 and 4

Singular plural

(a) 7 eke - 8 eβi-

12 aka - 8 eβi-

12 aka - 14 oβo-

(b) 5 eri- 6 ama-

skirtsrsquo

class 6

analyzed

Input bo-ranketi

a bo- ranketi

b ranketi

c blanket

(Kigali)

augmentation

Input gabana

a) o-mo-gaβana

b) mo-gaβana

c) gaβana

CHAPTER FIVE

51 Summary

that of English

English

Optimality Theory

prefixation

in nativization

52 Conclusions

53 Recommendations

nativization

word nativization

handle loans fully

REFERENCES

Harper and Row

Gruyter

press PP 270-292

MIT press

of Linguistics

Routledge

Education

Academic Press

OxfordOUP

2295-2310

MA MIT Press

contacts

University Press

University press

Oxford Blacwel

of Nairobi 2013)

14 6-18

sciences 2(8) 159-168

annual report

London)

productions

Berlin Routledge

Longman

Optimality Archive

linguisticsrdquo

1046-1078

Blakwell publishers

University press

and Co Ltd

Mounton deGruyer

Lingua116 950-975

APPENDICES

9 Knowlede of

Domain of the nouns

Questions

vie for)

9 LEGAL AFFAIRS

12BUSINESS TRADE

13 CLOTHING

Appendix V Raw data

- -soda sɒdǝ

SOURCE WORD FORM

deminution

1 Sarama

2 Nyagenke

3 Ikaraancha

4 Esuguta

5 Nyakoria

6 Nyankarankania

7 Nyantaro

8 Nyagaachi

9 Motagara

10 Mariba A

11 Mariba B

12 Chumura

13 Mosobeti

14 Getukora

15 Enchoro

16 Ikarancha

17 Nyando

07-Nov-2017 10 35AM (UTC + 0300)

DECLARATION

COPYRIGHT

ABSTRACT

DEDICATION

ACKNOWLEDGEMENT

TABLE OF CONTENTS

ABBREVIATIONS

SYMBOLS

CHAPTER ONE

INTRODUCTION

CHAPTER TWO

20 Introduction

CHAPTER THREE

30 Introduction

31 Research design

32 Research site

36 Data analysis

CHAPTER FOUR

40 Introduction

41522 The Prefix

CHAPTER FIVE

51 Summary

52 Conclusions

53 Recommendations

REFERENCES

APPENDICES

Appendix V Raw data

dishonesty

APPROACH

degree

(Candidate)

(Supervisor 1)

(Supervisor 2)

COPYRIGHT

ABSTRACT

DEDICATION

ACKNOWLEDGEMENT

deep indebtness

institutions

TABLE OF CONTENTS

DECLARATIONii

COPYRIGHTv

ABSTRACTvi

DEDICATIONvii

ACKNOWLEDGEMENTviii

TABLE OF CONTENTSx

CHAPTER ONE

INTRODUCTION1

CHAPTER TWO

20 Introduction12

CHAPTER THREE

30 Introduction40

32 Research site41

36 Data analysis43

CHAPTER FOUR

40 Introduction46

41522 The Prefix125

obo-265

CHAPTER FIVE

51 Summary276

52 Conclusions281

REFERENCES285

APPENDICES307

CHARTS

ABBREVIATIONS

AFX affix

APP applicative

C consonant

EVAL evaluator

FV final vowel

GEN generator

IND indicative mood

NC nasal Consonant

OM object Marker

P nominal Person

PL plural

SG singular

SM subject Marker

RT root of a word

TNS tense

V vowel

HTS high tone spans

correspondent tone

SYMBOLS

σ Syllable node

Optimal candidate

Fatal violation

Phonology

TABLES

TABLEAUX

CHARTS Page

FIGURES Page

xxviii

CHAPTER ONE

INTRODUCTION

spoken in Kenya

language

EkeGusii undergo

English nouns

into EkeGusii

this study

recipient language

CHAPTER TWO

20 Introduction

this study

phonology

is concerned

(Keith1997105)

EkeGusii

language

characteristics

Input GEN

Source Kager (1999)

chellip

gtgt gtgt

Output

correspondent tone

in tableau (1)

a) intip

b) skip

generative theories

are novel

constraints

different languages

CHAPTER THREE

30 Introduction

31 Research design

32 Research site

as follows

where it occurs

36 Data analysis

INPUT GEN OUTPUT

CHAPTER FOUR

40 Introduction

Time (s)0 0136

quency (

5000 ε

Time (s)0 007152

5000sound ee

Time (s)0 03045

Time (s)0 008388

quency (

5000sound ou

Time (s)0 007639

5000sounnd i

Time (s)0 1189

0594285714u

Time (s)0 007639

5000sounnd ii

Time (s)0 0136

ency (

5000ͻ

2500 2000 1500 1000 500

having a low F2)

4) English vowels

a) pure vowels

i heed feel bead

ɪ hid fill bid

e head fell bed

ᴂ had bad mad

ɑ hard bard par

ɒ hod bod

ʊ hood full

ʌ but cut hut

ǝ accept father

b) Dipthongs

eɪ fail bayed pay

ͻɪ foil boy

ǝʊ hoe foal

ɪǝ beard beer

ʊǝ poor

c) Triphthongs

ǝʊǝ mower slower

OrsquoConnor 1967)

Vowel Description

ᴂ ε

ͻ ͻͻ

languages

Table (2) above

Vowel F1 F2

i 540 2450e 730 2250ͻ 750 1250u 520 1100

Vowel F1 F2

i 280 2250e 400 1920ͻ 590 850u 310 890

DIPHTHONG

centring closing

(Ongarora 1996)

b) V1 V2 V3

c) V1 V2 V3

+HIGH])

Continuant β s γ

Obstruent [b] t k

Affricate c

Nasal m n ntilde ɳ

Glide w y

[Cw] [Cy]

Continuant ɸ s γ

(tril) r

Obstruent [b] t k

[d] [g]

Affricate [dʒ]

Nasal m n ɳ ŋ

Approximant [w] j

these processes

below demonstrates

[m b w a t e]

Output [wata]

Output sjeka

McCarthy (20079)

consonant glide [j]

c vcv c cvcv

clusters

a) rwana σ σ σ

C V V C V

rw u a n a

b) rwana σ σ

cc v c v

rw a n a

sought to answer

enda e nda

V NC V

e nd a

CVV CV C CVVC

node as figure (11)

σ σ σ σ

m o n t o m o n t o

(200290)

n-to- taatint -ɛ

F-1p-fetch-FV

[ntotaatinte]

Fndash 2p1-fetch-fv

[mmotaatinte]

c) in-a-taatint-e

F-3p-fetch-fv

[mbataatinte]

Suffixation -ee

Prefixation n-e-

figure (12)

[word [+Nasal]

Suffixation -γore

Prefixation n-γore

Output [ŋgore]

N C V C V

ŋ g o r e

CC V C V

ng o r e

syllables

i) ink rarr iŋk

cluster

X (C)CV(C) Dakota

XO CV(C)(C) Klamath

X CV(C) TemierR X

O _ (C)VC Arabela

X CV Senufo

X (C)V(C) Turkish

C V C V

γ e t ii

i) omote

vc vcv σ σ σ

v c v c v

o m o t e ii) a aa

vv σ σ

in this position

respectively

vowel word

i e letter ltegt

ǝ a ɑ are

ͻ ɜ err

eɪ a letter ltagt

aɪ i letter ltigt

ǝʊ owe

ɪǝ ear

eǝ air

others do not

Cruttenden (2011)

m+ j mew

n+ j new

l+ j lure leau

v+j view

h + j hew

C V V C V

γ a k a -γaaka

C V C V

γ a k a -γaka

L L H L

k כ bεlε mbu mba

H H HH

observation

H L H L H H L HLH

tone (HL)

shown in (38)

wik weak kɒntrʌkt

ɒdǝlɪ orderly

Singular Plural

meaning

(a) (b)

i) omot e σ

vc vc v σ σ σ

v c v c v

o m o t e

- mo- nto -βa- nto

-tata -βa- tata

rule as in (7)

vowel e

41522 The Prefix

previous section

aug- 3psg- root

aug-3pl- root

aug- 3psg- root

aug- 3pl- root

(i) Class 3

form form form

ii) Class 4

Class 9 e-

Class 10 chin-

Class 9a en-

Class 10a chin-

in (52)

Singular plural

(a) 7 eke - 8 eβi-

12 akas - 8 eβi-

12 aka- 14 oβo-

(b) 5 eri- 6 ama-

a) Diminution

a- ka- mo- onto

aug- 12- 1 - person

[akamoonto]

[akamoonto aγake]

[ekemoonto eγeke]

c) abusive

d) pejorative

phonology

(2001)

a [krɪsmǝs]

b [kirisimasi]

a [krɪsmǝs]

b[kirisimasi]

fatal violation

rounded low]

a [intɜt]

b esati

a [intɜt]

b [esati]

a [glɑs]

b [ekerasi]

a [glɑs]

b [ekerasi]

produce

a [sɪneɪt]

b [eseneti]

a [sɪneɪt]

b [eseneti]

ǝ [a]

a [lǝʊkeɪintn]

b [erokeseni]

a [lǝʊkeɪintn]

b [erokeseni]

English [aɪ]

a [waɪǝ]

b [waja]

a [waɪǝ]

b [waja]

a) frɪdʒ

b) eβuritinti

a) frɪdʒ

b) eβuritinti

a) skafu

b) sikava

minimal if at all

lsquothermosrsquo

a) Өɜməs

b) tamosi

c) ethɜməs

a) Өamosi

b) tamosi

c) ethɜməs

[filimu]

a) claeligs

b) ekerasi

c) ekelasi

a) sukuru

b) sukulu

in (85)

a) zirəʊ

b) esiro

a) esiro

b) zirəʊ

a) intɜt

b) esati

a) intɜt

b) esati

a) baeligg

b) epaγi

a) baeligg

b) epaγi

O N O N

C V C V

m a m a

N O N O N

V C V C V

o m o t e

CC V CC

s t eɪ nz

Input mama

a) mama

b) a ama

c) maa

a) ǝʊ

b) tǝʊ

c) ǝʊǝ

a) teɪ

b) teɪk

c) teɪki

a) seɪn

b) siteɪnsɪ

c) steɪns

shown in (90)

English EkeGusii

σ σ σ

O N O N O N

CC V C V C V

s t ͻ s i t o

nativized form

a) sukuru

a) skuru

c) sukuru

EkeGusii

English EkeGusii

σ σ σ σ

O N C O N O N O N

CC V C C V C V C V

Samoan Sranan

English EkeGusii

σ σ σ

O V C O N O N

C N C C V C V

s u t s u t i 192

a) tͻtint

b) tͻtinti

a) tͻtint

b) tͻͻtinti

Input t-n-ak-ol

a) t-n-ak-ol

b) t i nagͻl

a) noŋkomati

b) no-n-koma-i

illustrated by (96)

L H L H

L L H L H

L L H L

L H L H

c) Dholuo

Input blaeligŋkɪt

Input oo-raŋget-i

FLOAT MAX-IO

IDENT-IO

DEP-IO SPEC

a) okokoro

b) okoγoro

(transfer)

a) baeligŋk

b) eŋgi

c) e - ki - ar - a d) e - bi - ar- a

the plural form

11 boy

finger

ii) |r| rarr [d]

shown in (107)

Output [kambi]

a kambi

b kani

c kanbi

defricativization

are both bilabials)

a) eteγisi

b) tᴂksi

[γ] as the output

sub-section

in tableau (42)

a ɸaranda

b vǝrᴂndǝ

i) drink

(ii) veranda

Input e-+βaranr-a

English EkeGusii

veranda baranda

tableau (43)

a ɸaranda (unit)

UF SF Gloss

a ɛntɛrɛa

b draɪvǝ

follows

a nkooko

b e-nkooko

a e-nkooko

b nk-ooko

b boy-s

a boy-s

Input blanket-s

(50) as follows

a blanket

b ama- ranketi

c blanket-s

nominal class

form as in (118)

objects

below show

Input tractor-s

analysis

L RT R)) STRPRES

a prefix

Input o- mo- te

aug 33PSG tree

output [omote]

b record-s

c s-record

at syllable margins

in tableau (55)

Input recordrekͻ d

a rekͻd-s

b s-rekͻd

c rekͻd

933PSG- record

a e-rekoti

b record

c record-s

highly ranked

103PPL record

a chi-rekoti

b record

c record-chi

violations

are 12 34 14 6 7 8 and 12

a o-mo- chiibu

b chief

c chief-s

Input chief

a o-mo- chiibu

b chief

c chief-s

prefix

denote plurals

classes 3 and 4

Singular plural

(a) 7 eke - 8 eβi-

12 aka - 8 eβi-

12 aka - 14 oβo-

(b) 5 eri- 6 ama-

skirtsrsquo

class 6

analyzed

Input bo-ranketi

a bo- ranketi

b ranketi

c blanket

(Kigali)

augmentation

Input gabana

a) o-mo-gaβana

b) mo-gaβana

c) gaβana

CHAPTER FIVE

51 Summary

that of English

English

Optimality Theory

prefixation

in nativization

52 Conclusions

53 Recommendations

nativization

word nativization

handle loans fully

REFERENCES

Harper and Row

Gruyter

press PP 270-292

MIT press

of Linguistics

Routledge

Education

Academic Press

OxfordOUP

2295-2310

MA MIT Press

contacts

University Press

University press

Oxford Blacwel

of Nairobi 2013)

14 6-18

sciences 2(8) 159-168

annual report

London)

productions

Berlin Routledge

Longman

Optimality Archive

linguisticsrdquo

1046-1078

Blakwell publishers

University press

and Co Ltd

Mounton deGruyer

Lingua116 950-975

APPENDICES

9 Knowlede of

Domain of the nouns

Questions

vie for)

9 LEGAL AFFAIRS

12BUSINESS TRADE

13 CLOTHING

Appendix V Raw data

- -soda sɒdǝ

SOURCE WORD FORM

deminution

1 Sarama

2 Nyagenke

3 Ikaraancha

4 Esuguta

5 Nyakoria

6 Nyankarankania

7 Nyantaro

8 Nyagaachi

9 Motagara

10 Mariba A

11 Mariba B

12 Chumura

13 Mosobeti

14 Getukora

15 Enchoro

16 Ikarancha

17 Nyando

07-Nov-2017 10 35AM (UTC + 0300)

DECLARATION

COPYRIGHT

ABSTRACT

DEDICATION

ACKNOWLEDGEMENT

TABLE OF CONTENTS

ABBREVIATIONS

SYMBOLS

CHAPTER ONE

INTRODUCTION

CHAPTER TWO

20 Introduction

CHAPTER THREE

30 Introduction

31 Research design

32 Research site

36 Data analysis

CHAPTER FOUR

40 Introduction

41522 The Prefix

CHAPTER FIVE

51 Summary

52 Conclusions

53 Recommendations

REFERENCES

APPENDICES

Appendix V Raw data

APPROACH

degree

(Candidate)

(Supervisor 1)

(Supervisor 2)

COPYRIGHT

ABSTRACT

DEDICATION

ACKNOWLEDGEMENT

deep indebtness

institutions

TABLE OF CONTENTS

DECLARATIONii

COPYRIGHTv

ABSTRACTvi

DEDICATIONvii

ACKNOWLEDGEMENTviii

TABLE OF CONTENTSx

CHAPTER ONE

INTRODUCTION1

CHAPTER TWO

20 Introduction12

CHAPTER THREE

30 Introduction40

32 Research site41

36 Data analysis43

CHAPTER FOUR

40 Introduction46

41522 The Prefix125

obo-265

CHAPTER FIVE

51 Summary276

52 Conclusions281

REFERENCES285

APPENDICES307

CHARTS

ABBREVIATIONS

AFX affix

APP applicative

C consonant

EVAL evaluator

FV final vowel

GEN generator

IND indicative mood

NC nasal Consonant

OM object Marker

P nominal Person

PL plural

SG singular

SM subject Marker

RT root of a word

TNS tense

V vowel

HTS high tone spans

correspondent tone

SYMBOLS

σ Syllable node

Optimal candidate

Fatal violation

Phonology

TABLES

TABLEAUX

CHARTS Page

FIGURES Page

xxviii

CHAPTER ONE

INTRODUCTION

spoken in Kenya

language

EkeGusii undergo

English nouns

into EkeGusii

this study

recipient language

CHAPTER TWO

20 Introduction

this study

phonology

is concerned

(Keith1997105)

EkeGusii

language

characteristics

Input GEN

Source Kager (1999)

chellip

gtgt gtgt

Output

correspondent tone

in tableau (1)

a) intip

b) skip

generative theories

are novel

constraints

different languages

CHAPTER THREE

30 Introduction

31 Research design

32 Research site

as follows

where it occurs

36 Data analysis

INPUT GEN OUTPUT

CHAPTER FOUR

40 Introduction

Time (s)0 0136

quency (

5000 ε

Time (s)0 007152

5000sound ee

Time (s)0 03045

Time (s)0 008388

quency (

5000sound ou

Time (s)0 007639

5000sounnd i

Time (s)0 1189

0594285714u

Time (s)0 007639

5000sounnd ii

Time (s)0 0136

ency (

5000ͻ

2500 2000 1500 1000 500

having a low F2)

4) English vowels

a) pure vowels

i heed feel bead

ɪ hid fill bid

e head fell bed

ᴂ had bad mad

ɑ hard bard par

ɒ hod bod

ʊ hood full

ʌ but cut hut

ǝ accept father

b) Dipthongs

eɪ fail bayed pay

ͻɪ foil boy

ǝʊ hoe foal

ɪǝ beard beer

ʊǝ poor

c) Triphthongs

ǝʊǝ mower slower

OrsquoConnor 1967)

Vowel Description

ᴂ ε

ͻ ͻͻ

languages

Table (2) above

Vowel F1 F2

i 540 2450e 730 2250ͻ 750 1250u 520 1100

Vowel F1 F2

i 280 2250e 400 1920ͻ 590 850u 310 890

DIPHTHONG

centring closing

(Ongarora 1996)

b) V1 V2 V3

c) V1 V2 V3

+HIGH])

Continuant β s γ

Obstruent [b] t k

Affricate c

Nasal m n ntilde ɳ

Glide w y

[Cw] [Cy]

Continuant ɸ s γ

(tril) r

Obstruent [b] t k

[d] [g]

Affricate [dʒ]

Nasal m n ɳ ŋ

Approximant [w] j

these processes

below demonstrates

[m b w a t e]

Output [wata]

Output sjeka

McCarthy (20079)

consonant glide [j]

c vcv c cvcv

clusters

a) rwana σ σ σ

C V V C V

rw u a n a

b) rwana σ σ

cc v c v

rw a n a

sought to answer

enda e nda

V NC V

e nd a

CVV CV C CVVC

node as figure (11)

σ σ σ σ

m o n t o m o n t o

(200290)

n-to- taatint -ɛ

F-1p-fetch-FV

[ntotaatinte]

Fndash 2p1-fetch-fv

[mmotaatinte]

c) in-a-taatint-e

F-3p-fetch-fv

[mbataatinte]

Suffixation -ee

Prefixation n-e-

figure (12)

[word [+Nasal]

Suffixation -γore

Prefixation n-γore

Output [ŋgore]

N C V C V

ŋ g o r e

CC V C V

ng o r e

syllables

i) ink rarr iŋk

cluster

X (C)CV(C) Dakota

XO CV(C)(C) Klamath

X CV(C) TemierR X

O _ (C)VC Arabela

X CV Senufo

X (C)V(C) Turkish

C V C V

γ e t ii

i) omote

vc vcv σ σ σ

v c v c v

o m o t e ii) a aa

vv σ σ

in this position

respectively

vowel word

i e letter ltegt

ǝ a ɑ are

ͻ ɜ err

eɪ a letter ltagt

aɪ i letter ltigt

ǝʊ owe

ɪǝ ear

eǝ air

others do not

Cruttenden (2011)

m+ j mew

n+ j new

l+ j lure leau

v+j view

h + j hew

C V V C V

γ a k a -γaaka

C V C V

γ a k a -γaka

L L H L

k כ bεlε mbu mba

H H HH

observation

H L H L H H L HLH

tone (HL)

shown in (38)

wik weak kɒntrʌkt

ɒdǝlɪ orderly

Singular Plural

meaning

(a) (b)

i) omot e σ

vc vc v σ σ σ

v c v c v

o m o t e

- mo- nto -βa- nto

-tata -βa- tata

rule as in (7)

vowel e

41522 The Prefix

previous section

aug- 3psg- root

aug-3pl- root

aug- 3psg- root

aug- 3pl- root

(i) Class 3

form form form

ii) Class 4

Class 9 e-

Class 10 chin-

Class 9a en-

Class 10a chin-

in (52)

Singular plural

(a) 7 eke - 8 eβi-

12 akas - 8 eβi-

12 aka- 14 oβo-

(b) 5 eri- 6 ama-

a) Diminution

a- ka- mo- onto

aug- 12- 1 - person

[akamoonto]

[akamoonto aγake]

[ekemoonto eγeke]

c) abusive

d) pejorative

phonology

(2001)

a [krɪsmǝs]

b [kirisimasi]

a [krɪsmǝs]

b[kirisimasi]

fatal violation

rounded low]

a [intɜt]

b esati

a [intɜt]

b [esati]

a [glɑs]

b [ekerasi]

a [glɑs]

b [ekerasi]

produce

a [sɪneɪt]

b [eseneti]

a [sɪneɪt]

b [eseneti]

ǝ [a]

a [lǝʊkeɪintn]

b [erokeseni]

a [lǝʊkeɪintn]

b [erokeseni]

English [aɪ]

a [waɪǝ]

b [waja]

a [waɪǝ]

b [waja]

a) frɪdʒ

b) eβuritinti

a) frɪdʒ

b) eβuritinti

a) skafu

b) sikava

minimal if at all

lsquothermosrsquo

a) Өɜməs

b) tamosi

c) ethɜməs

a) Өamosi

b) tamosi

c) ethɜməs

[filimu]

a) claeligs

b) ekerasi

c) ekelasi

a) sukuru

b) sukulu

in (85)

a) zirəʊ

b) esiro

a) esiro

b) zirəʊ

a) intɜt

b) esati

a) intɜt

b) esati

a) baeligg

b) epaγi

a) baeligg

b) epaγi

O N O N

C V C V

m a m a

N O N O N

V C V C V

o m o t e

CC V CC

s t eɪ nz

Input mama

a) mama

b) a ama

c) maa

a) ǝʊ

b) tǝʊ

c) ǝʊǝ

a) teɪ

b) teɪk

c) teɪki

a) seɪn

b) siteɪnsɪ

c) steɪns

shown in (90)

English EkeGusii

σ σ σ

O N O N O N

CC V C V C V

s t ͻ s i t o

nativized form

a) sukuru

a) skuru

c) sukuru

EkeGusii

English EkeGusii

σ σ σ σ

O N C O N O N O N

CC V C C V C V C V

Samoan Sranan

English EkeGusii

σ σ σ

O V C O N O N

C N C C V C V

s u t s u t i 192

a) tͻtint

b) tͻtinti

a) tͻtint

b) tͻͻtinti

Input t-n-ak-ol

a) t-n-ak-ol

b) t i nagͻl

a) noŋkomati

b) no-n-koma-i

illustrated by (96)

L H L H

L L H L H

L L H L

L H L H

c) Dholuo

Input blaeligŋkɪt

Input oo-raŋget-i

FLOAT MAX-IO

IDENT-IO

DEP-IO SPEC

a) okokoro

b) okoγoro

(transfer)

a) baeligŋk

b) eŋgi

c) e - ki - ar - a d) e - bi - ar- a

the plural form

11 boy

finger

ii) |r| rarr [d]

shown in (107)

Output [kambi]

a kambi

b kani

c kanbi

defricativization

are both bilabials)

a) eteγisi

b) tᴂksi

[γ] as the output

sub-section

in tableau (42)

a ɸaranda

b vǝrᴂndǝ

i) drink

(ii) veranda

Input e-+βaranr-a

English EkeGusii

veranda baranda

tableau (43)

a ɸaranda (unit)

UF SF Gloss

a ɛntɛrɛa

b draɪvǝ

follows

a nkooko

b e-nkooko

a e-nkooko

b nk-ooko

b boy-s

a boy-s

Input blanket-s

(50) as follows

a blanket

b ama- ranketi

c blanket-s

nominal class

form as in (118)

objects

below show

Input tractor-s

analysis

L RT R)) STRPRES

a prefix

Input o- mo- te

aug 33PSG tree

output [omote]

b record-s

c s-record

at syllable margins

in tableau (55)

Input recordrekͻ d

a rekͻd-s

b s-rekͻd

c rekͻd

933PSG- record

a e-rekoti

b record

c record-s

highly ranked

103PPL record

a chi-rekoti

b record

c record-chi

violations

are 12 34 14 6 7 8 and 12

a o-mo- chiibu

b chief

c chief-s

Input chief

a o-mo- chiibu

b chief

c chief-s

prefix

denote plurals

classes 3 and 4

Singular plural

(a) 7 eke - 8 eβi-

12 aka - 8 eβi-

12 aka - 14 oβo-

(b) 5 eri- 6 ama-

skirtsrsquo

class 6

analyzed

Input bo-ranketi

a bo- ranketi

b ranketi

c blanket

(Kigali)

augmentation

Input gabana

a) o-mo-gaβana

b) mo-gaβana

c) gaβana

CHAPTER FIVE

51 Summary

that of English

English

Optimality Theory

prefixation

in nativization

52 Conclusions

53 Recommendations

nativization

word nativization

handle loans fully

REFERENCES

Harper and Row

Gruyter

press PP 270-292

MIT press

of Linguistics

Routledge

Education

Academic Press

OxfordOUP

2295-2310

MA MIT Press

contacts

University Press

University press

Oxford Blacwel

of Nairobi 2013)

14 6-18

sciences 2(8) 159-168

annual report

London)

productions

Berlin Routledge

Longman

Optimality Archive

linguisticsrdquo

1046-1078

Blakwell publishers

University press

and Co Ltd

Mounton deGruyer

Lingua116 950-975

APPENDICES

9 Knowlede of

Domain of the nouns

Questions

vie for)

9 LEGAL AFFAIRS

12BUSINESS TRADE

13 CLOTHING

Appendix V Raw data

- -soda sɒdǝ

SOURCE WORD FORM

deminution

1 Sarama

2 Nyagenke

3 Ikaraancha

4 Esuguta

5 Nyakoria

6 Nyankarankania

7 Nyantaro

8 Nyagaachi

9 Motagara

10 Mariba A

11 Mariba B

12 Chumura

13 Mosobeti

14 Getukora

15 Enchoro

16 Ikarancha

17 Nyando

07-Nov-2017 10 35AM (UTC + 0300)

DECLARATION

COPYRIGHT

ABSTRACT

DEDICATION

ACKNOWLEDGEMENT

TABLE OF CONTENTS

ABBREVIATIONS

SYMBOLS

CHAPTER ONE

INTRODUCTION

CHAPTER TWO

20 Introduction

CHAPTER THREE

30 Introduction

31 Research design

32 Research site

36 Data analysis

CHAPTER FOUR

40 Introduction

41522 The Prefix

CHAPTER FIVE

51 Summary

52 Conclusions

53 Recommendations

REFERENCES

APPENDICES

Appendix V Raw data

Documents

PHONOLOGICAL AND MORPHOLOGICAL NATIVIZATION OF …