PHONOLOGICAL PROCESSES IN SELECTED NIGERIAN CHILDREN’S SPOKEN ENGLISH IN OYO AND LAGOS STATES, NIGERIA Y BY R BR A L I RONKE EUNICE ADESOYE (MATRIC. NO.: 143308)N B.A., M.A. (IbDadanA) A IB A T hO F esis in the Department of English SubmittYed to the Faculty of Arts in partial fulfilment of IT the requirements for the Degree of S DOCTOR OF PHILOSOPHY R of the E UNIVERSITY OF IBADAN NI V U OCTOBER, 2021 CERTIFICATION I certify that this work was carried out by Ronke Eunice ADESOYE (Matric. No.: 143308) in the Department of English, University of Ibadan, Nigeria under my supervision. RY RA _________________________ IB Supervisor L A.B. Sunday B.A., M.A. Ph.D. (IbadanN) Department of EnglAish, University of IbadaDn, Nigeria IB A F Y O T SI VE R UN I ii DEDICATION I dedicate this research to the One who moulds bones out of liquid comprehends the unspoken words of infants, and bestows on them the ability to acquire what adults learn RY A R LI B AN D IB A O F TY RS I E NI V U iii ACKWNOWLEDGEMENTS I thank God for the gift of life and strength to begin and complete this research. Without Him, I am nothing. I thank Him for the gift of invaluable people whom He used to help me achieve this feat. First on this long list of the gift of people is my supervisor and mentor – Dr AdeYsina Bukunmi Sunday. He has not only supervised me, he has fathered, mentored andR guided me since the first time I came to his awareness. God bless him immenselyA. I thank Dr Yewande Sunday for her pieces of advice and support. To me and theB resRt of Dr Sunday’s supervision group, she has indeed been a mother. LI I would also like to acknowledge the support and contributNion o f all the academic and non-academic staff members of the Department of EngAlish. I thank my dear teachers for providing the bedrock of knowledge from which DI could spring forth. They taught me the fundamentals for good character, discipline and knowledge. Furthermore, I thank the University Management for the magnanimouAs opportunity to carry out my research while enjoying the benefit of rebate as a univ eIrsBity staff. I also thank Professor Ayobami Kehinde for coming to the rescue wheneverF hitches came up in this journey. I thank my parents, Pastors Oluyemisi and Bamidele Adesoye, for encouraging me to further my education. ItY has proved beyond measure to be a worthwhile endeavour. Their guidance, prayers, IanTd advice have provided a strong pillar of support. I thank mRy deSar husband, Larry Eghosa Okhuosi, for his love, prayers, counsel and objectiEvity, and for not letting me slip into mediocrity. I thank him for loving me as Christ lovVes the Church. Our love will continue to flourish in Jesus’ name, Amen. NFIurthermore, I thank my daughter, Osarugue Inioluwa, for cooperating with me all through U the duration of this programme. She made it easy to combine the responsibilities of a mother, teacher and student. I love you. God bless you. I acknowledge my ‘parents-in-love’, Mr Stevenson and Dr Angela Okhuosi. I appreciate them for their concern and counsel about my thesis which showed quite clearly that they care about my progress. Also, I appreciate my siblings (in-love), Oluwagbenga, Richard, iv Sandra, Oluwaseun, Oluwamayowa, Kelly, George, Ernest, Stephanie and Jessie. I thank them for believing in, and trusting me to deliver ahead of this achievement. I also appreciate Humboldt-Universität, Berlin, for the four-month opportunity of studying there. Specifically, I thank Professors Remi Raji-Oyelade, Susanne Gehrmann, Tom Güldermann and Dr Christfried Naumann at the Institute of Asian and African StudieYs. I also acknowledge the learning opportunity with Edna Olondo, Nicholus Makanji,R Neema, Fidelis, Akoete, and a host of other scholars I met in Berlin. A I thank Mrs Opeyemi Ajibola for her encouragement, questions, and praRyers. Very many times, her strength was enough for the two of us. I appreciate Dr GaInBiu Bamgbose and my fellow supervisees for the tips, questions, advice, materials, pLosts, and guidance, I am grateful for these and much more. N I thank Folakemi Ademodi-Abiodun for her frienDdshAip, prayers, support, encouragement and counsel. She always had a solution to all Athe problems I came to her with and she also always had enough will and persistence foBr the two of us. I thank Temiloluwa Bodunde for lending me her support and encourage mIent, even from across the seven oceans. I thank everyone who has imOpacFted me directly or indirectly during the course of this research. TY I Adesoye, R.E. S October, 2021 VE R UN I v ABSTRACT Phonological processes constitute a veritable means to tracing language development, especially in children. Extant studies on Nigerian children’s phonological processes have examined errors and deviations, with little attention to language as an instrument for measuring children’s linguistic development. Therefore, this study was designed to examine children’s phonological processes and the constraints ranking responsibleY for them, with a view to tracing their linguistic development. Alan Prince and Paul Smolensky’s Optimality Theory was used as the framework, while the descriptive design was adopted. One hundred and two participants wereA purRposively selected. Seventy-five and twenty-five children from primary schools in Lagos and Oyo states, respectively, were selected because of their age range of four to sixR years, and they read a prepared text. The choice of the states was motivated by their proximal, cosmopolitan and multicultural features. Also, two children, named child A, aged onBe year-three months, and child B, aged four years-three months, were observed in th eiLr ho Imes in Oyo and Lagos states, respectively, for six months for the purpose of longitudinal observation. All utterances were audio-recorded. Data were subjected to dNescriptive statistics, perceptual and acoustic analyses. The phonological processes identified were substituAtion (28.8%), vowel strengthening (23.2%), monophthongisation (15.7%), deletionD (15.4%), assimilation (6.6%), gliding (4.3%) and yod coalescence (2.7%). UtterancAes were slow-paced, with an average of 4.8 minutes per participant, and phonemes were often singly produced. Constraints ranking favoured markedness over faithfulneBss constraints, such as *SCHWA >> αF, NODIPHTHONG >> MAX, *Ct#C >> *C OIMPLEX >> MAX and AGREE(PLACE) >> IDENT- IO. The participants’ linguistic deFvelopment was noticeable in the instantiations of their processes, which were similar to the ambient variety of Nigerian English. The instances were very intelligible and sigOnificantly manifested beyond word level. They were also functional for achieving Yjunc ture prosody, cluster reduction and gemination. However, non-adult instances, likeT morphophonemic deletions, persisted, showing that the participants had not fully attained the adult level of phonological processes. In the longitudinal data, child A acquireSd voIiced and labial consonants first, and codas suffered deletion more than onsets in monosyllables. By age two, child A had begun to produce polysyllables and closed syllables, aRnd deletion changed from whole syllables to only phonemes. By age five, child B’s proEcesses had begun to resemble adults’ and, more energy-demanding processes like epenthesis, voicing and vowel strengthening emerged. Tonalisation of English words and inIdVigenous interference occurred in their utterances. The spectrogram showed that the outset of acquisition with child A featured weaker energy, like in unaspirated plosives; Nhowever, energy increased and stabilised as the participant got older, as indicated in the U darker shades. The formant values of the participants’ vowels on the acoustic chart showed similarity to the cardinal vowel chart in terms of height and position of the tongue. Phonological processes in Nigerian children’s spoken English emerged through constraints reranking and increasingly become more like adults’ as the years pass by. Keywords: Linguistic development, Language acquisition, Nigerian children’s English Word count: 473 vi TABLE OF CONTENTS Title Page Title page i Certification ii Dedication iii Y Acknowledgements Riv Abstract A vi Table of contents R vii List of tables B xi List of figures LI xii List of tableaux xiii CHAPTER ONE: INTRODUCTION N 1.1 Background to the study D A 1 1.2 Statement of the problem 3 1.3 Aim and objectives A 4 1.4 Scope of the study IB 4 1.5 Significance of the study F 4 1.6 Summary O 5 CHAPTER TWO: LITYER ATURE REVIEW AND THEORETICAL FRAMEWORK T 2.0 PreambSle I 6 2.1 ConRceptual review 6 2.1.1 Language acquisition 6 2.1V.2 ELanguage acquisition studies 8 2I.1.3 Theories of language acquisition 10 N2.1.4 Logical problem of language acquisition (LPLA) 24 U 2.1.5 The dilemma of the “how” 25 2.1.6 Psycholinguistics 28 2.1.7 Stages of language acquisition 30 2.1.8 Phonological processes 39 2.2 Empirical literature on child language studies 44 vii 2.2.1 Prescriptive studies on child language 44 2.2.2 Descriptive studies on child language 46 2.2.3 Phonological processes 50 2.2.4 Appraisal of literature review 53 2.3 International Phonetic Alphabet 54 2.4 Theoretical framework: Optimality Theory (OT) 56 Y 2.4.1 GEN(erator) R58 2.4.2 CON(straint) A 59 2.4.3 EVAL(uator) BR 61 2.4.4 Tableau I 62 2.4.5 Optimality Theory and phonological processes L 66 2.5 Summary N 67 CHAPTER THREE: METHODOLOGY A 3.0 Preamble D 68 3.1 Research design A 68 3.2 Sampling technique IB 68 3.3 Sample size F 68 3.4 Research instrument O 69 3.5 Ethical approval 69 3.6 Method of daTta cYollection 70 3.7 Method of dIata analysis 70 3.8 SummaSry 71 CHAPTER FOUR: DATA ANALYSIS AND DISCUSSION 4.0V EPreamble 72 4I.1 Socio-demographic description of the participants 72 N4.2 Phonological processes 76 U 4.2.1 Coalescence 76 4.2.2 Substitution 79 4.2.3 Monophthongisation 83 4.2.4 Assimilation 87 4.2.5 Deletion 91 viii 4.2.6 Vowel strengthening 95 4.2.7 Gliding 98 4.3 Constraints ranking for processes 98 4.3.1 Constraints ranking for coalescence 98 4.3.1.1 Realisation of second year 99 4.3.1.2 Coalescence of /e/ and /i/ 101Y 4.3.1.3 Representative acoustic analysis of coalescence in /ɛ/ R104 4.3.2 Constraints ranking for substitution R A 110 4.3.2.1 Realisation of birthday B 110 4.3.2.2 Realisation of with I 113 4.3.2.3 Realisation of the L 115 4.3.2.4 Realisation of father N 118 4.3.2.5 Realisation of thing A 120 4.3.2.6 Representative acoustic analysis of substitDuti on 123 4.3.3 Constraints ranking for MonophthongAisation 127 4.3.3.1 Realisation of I IB 127 4.3.3.2 Realisation of gave 129 4.3.3.3 Realisation of rose F 131 4.3.3.4 Realisation of where O 133 4.3.3.5 Realisation of maYy 135 4.3.3.6 RepresentatIivTe acoustic analysis of monophthongisation 137 4.3.4 ConstraSints ranking for Assimilation 141 4.3.4.1 ReaRlisation of like blue 141 4.3V.4.2E Realisation of have password 143 4I.3.4.3 Realisation of give me 146 N4.3.4.4 Representative acoustic analysis of assimilation 148 U 4.3.5 Constraints ranking for deletion 150 4.3.5.1 Realisation of Textbooks 150 4.3.5.2 Realisation of Friends 152 4.3.5.3 Realisation of want 154 4.3.5.4 Realisation of come 156 ix 4.3.5.5 Representative acoustic illustration of deletion 158 4.3.6 Constraints ranking for vowel strengthening 160 4.3.6.1 Realisation of Cut 162 4.3.6.2 Realisation of Our 162 4.3.6.3 Realisation of Away 164Y 4.3.6.4 Realisation of birthday R166 4.3.6.6 Realisation of another A 168 4.3.6.7 Representative acoustic analysis of vowel strengthening 171 4.3.7 Gliding BR 174 4.4 Processes exclusive to the longitudinal data I 176 4.4.1 Epenthesis L 176 4.4.1.1 Epenthesis in “Press zero” N 177 4.4.1.2 Representative acoustic analysis of epenthesis A 179 4.4.2 Devoicing D 180 4.4.3. Stopping A 184 4.5 Summary IB 188 CHAPTER FIVE: SUMMARY,F CONCLUSION AND RECOMMENDATIONS 189 5.0 Introduction O 189 5.1 Summary of Tthe Ystudy 189 5.2 Summary oIf findings 189 5.2.1 PhonoloSgical processes in the children’s language use 189 5.2.2 ConRstraints ranking for the processes 190 5.2V.3 ELanguage development 191 5I.3 Conclusion 192 N5.4 Recommendations 192 U 5.5 Limitations of the study 193 5.6 Contributions to knowledge 193 5.7 Suggestions for further studies 194 References 196 Appendix 204 x LIST OF TABLES Title Page Table 2.1 History of child language studies 9 Table 4.1 Age of the respondents 73 Table 4.2 Distribution of the participants between the states 75 Y Table 4.3 Realisation of coalescence in four to six year olds R77 Table 4.4 Realisation of substitution in four to six year olds A 81 Table 4.5 Realisation of monophthongisation in four to six year olds R 84 Table 4.6 Realisation of assimilation in four to six year olds I B 88 Table 4.7 Realisation of deletion in four to six year olds L 92 Table 4.8 Realisation of vowel strengthening in four to sixN yea r olds 96 AD A F IB O TYI S ER IV UN xi LIST OF FIGURES Title Page Fig. 2.1 Relationship between external stimuli and internal knowledge 16 Fig. 2.2 Factor interactions in language acquisition 22 Fig. 2.3 The International Phonetic Alphabet 55 Y Fig. 2.4 A schematic grammar at work R57 Fig. 4.1 Acoustic illustration of /e/ in /geit/ by a six-year-old A 105 Fig. 4.2 Acoustic illustration of /i/ in /geit/ by a six-year-old R 107 Fig. 4.3 Acoustic vowel chart for coalescence in /ɛ/ by a four-year-oIldB 109 Fig. 4.4 Acoustic illustration of substitution in /wit/ “with” by a Lfive-year-old 124 Fig. 4.5 Acoustic illustration of substitution in /di/ “the” by a f ive-year-old 126 Fig. 4.6 Acoustic analysis of monophthongisation in /roz/ Nby a five-year-old 138 Fig. 4.7 Acoustic vowel chart for /o/ by a five-yeaDr-oldA 140 Fig. 4.8 Acoustic illustration of assimilation iAn /laigblu/ by a five-year-old 149 Fig. 4.9 Acoustic analysis of deletion in “second” /'sekɔn/ by a four-year-old 159 Fig. 4.10 Acoustic analysis of vowel s trIenBgthening in /kɔ:t/ by a four-year-old 171 Fig. 4.11 Acoustic vowel chart of Fvowel strengthening in /kɔ:t/ cut 173 Fig. 4.12 Acoustic analysis ofO epenthesis in /preziziro/ by a four-year-old 179 Fig. 4.13 Acoustic analyYsis o f devoicing in “dog” /dɔ:k/ by a one-year-old 183 Fig. 4.14 Acoustic IanTalysis of stopping in “Sit” /ti:/ by a one-year-old 187 RS IV E UN xii LIST OF TABLEAUX Title Page Tableau 2.1 Comparative tableau 63 Tableau 2.2 Violation tableau 63 Tableau 2.3 Illustrative tableau 65 Tableau 4.1 Emergence of /sekɔnjie/ “second year” R100Y Tableau 4.2 Emergence of /ɛ/ “A” 103 Tableau 4.3 Emergence of /bedɛ/ “birthday” R A 111 Tableau 4.4 Emergence of /wit/ “with” B 114 Tableau 4.5 Emergence of /di/ “the” LI 117 Tableau 4.6 Emergence of /fada/ “father” 119 Tableau 4.7 Emergence of /tin/ “thing” N 121 Tableau 4.8 Emergence of /ai/ “I” A 128 Tableau 4.9 Emergence of /gɛv/ “gave” D 130 Tableau 4.10 Emergence of /roz/ “rose” A 132 Tableau 4.11 Emergence of /we/ “whereI”B 134 Tableau 4.12 Emergence of /mɛ/ “may ” 136 Tableau 4.13 Emergence of /lOaigbFlu/ “Like blue” 142 Tableau 4.14 Emergence of /haf/ “Have” 144 Tableau 4.15 EmergTenceY of /mimi/ 147 Tableau 4.16 EmerIgence of /tezbuk/ “textbooks” 151 Tableau 4.17 ESmergence of /frend/ 153 Tableau 4.R18 Emergence of /wɔn/ want 155 TableaEu 4.19 Emergence of /kɔ/ come 157 TIaVbleau 4.20 Emergence of /kɔt/ “cut” 161 NTableau 4.21 Emergence of /awa/ “our” 163 U Tableau 4.22 Emergence of /awɛ/ “away” 165 Tableau 4.23 Emergence of /bedɛ/ “birthday” 167 Tableau 4.24 Emergence of /anɔda/ “another” 169 Tableau 4.25 Emergence of /awa/ 175 Tableau 4.26 Emergence of /preziziro/ 177 xiii Tableau 4.27 Emergence of /dͻ:k/ 181 Tableau 4.28 Emergence of /ti/ “Sit” 185 Y RA R B LI AN AD IB OF TY SI VE R I UN xiv CHAPTER ONE Y INTRODUCTION R 1.1 Background to the study A Children often undergo rapid development, especially in the eaBrliesRt years of life. Understanding how this development occurs can be very instrumentaIl to identifying stunted growth very early. Child development entails the different cha nLges that children undergo between infancy and teen age. These changes can be descrNibed using at least five domains, and these are cognitive skills, language skills, motoAr skills, socio-emotional skills and executive functions. The cognitive skills relate to tDhought processes and memory; the socio- emotional skills are connected with relationsAhips, self-identity and empathy. The motor skill involves when and how well the chIiBld learns to sit, walk, run, jump, climb and so on. Executive functions refer to how a ch ild deals with a novel challenge. The language skill involves children’s production andF comprehension of words (Fernald, Kariger, Engle and Raikes, 2009). O In developed countrTies,Y these five domains are jointly used to follow the progress of children’s developIment. However, in developing countries, Fernald, et al. (2009) submit that the heightS of children has been the major yardstick for this description. Wells and StackhouseR (2016) also note that there is a dearth of child language studies that focus on languaEge skills as a domain of development in Africa. This claim is further corroborated bIy VAkanbi and Anyarsor (2014) who add that growth monitoring has been advocated by Nthe World Health Organization (WHO) and United Nations International Children U Emergency Funds (UNICEF) because of the high mortality rate of under-five-year-old children in the African region. Rather than tackle this problem holistically using the different domains, many Nigerian hospitals and clinics focus mainly on motor skills and physical development while the internal manifestations are neglected (Popoola and Adeoti 2016). Hence, other domains of 1 development receive less attention. Although this is not what obtains in the country, Fernald, Kariger, Engle and Raikes (2009) posit that these multiple domains need to be jointly investigated to capture a wider spectrum of children’s development and abilities. For instance, if a child experienced a setback in language development, such child would be negatively affected in the cognition and socio-emotional domains also, because these other domains require some level of language skill. Hence, the cognition and soYcio- emotional domains of such child would need to be examined too. AR About the language domain specifically, Bowen (2011) states that chilRdren’s language development is further broken down into four major aspects whicBh are intelligibility, phonological development, phonetic development and pLhonIotactic development. Intelligibility is the comprehension of what the children say as well as their comprehension of what is said to them. Phonological development refNers to how speech sounds are mentally stored in the child’s brain, how they are produAced as well as the processes between these two phases. Phonetic development is the aDctual articulation of these sounds while phonotactic development involves the chiBld’sA formation of words in the ambient language. However, some of Bowen’s (2011) di ffIerentiations within a child’s language development seem unnecessary as elements aFnd processes of phonotactic development are already embedded within phonolog icOal development. Bowen (2011) adds that while children undergo these four asYpects of language development, they engage in phonological processes – a proceIsTs that children subject their speech to from infancy to about five years of age (Brown, 1973; Levelt, 2013). Wells and RStacSkhouse (2016) opine that children under five years in low- and middle- incomeE countries are not manifesting the expected features of growth, cognition or socio- eIcoVnomic development. Nigeria, alongside many other African countries, falls within this Ndescription. As a form of response to this challenge, a number of scholars, like Rossington U (1981), Akanbi and Anyarsor (2014), Popoola and Adeoti (2016), Eze, Oguonu, Ojinnaka and Ibe (2017), and Jimoh, Anyiam and Yakubu (2018) have carried out child development-related research, from nutritional and medical perspectives, leaving the language domain lagging greatly in research and contribution to the description of child development in Nigeria, particularly in the area of phonology. 2 This is not to say that there are no phonological studies on Nigerian children. There are such studies as Akpan (2006), Ambrose (2010), Alerechi (2011), Onwubiko (2011), Dopemu (2015), Ibrahim and Ibrahim (2016), Adeniyi and Adeniyi (2017), Olarewaju (2018), Ademola-Adeoye (2019), Olarewaju and Sunday (2020), and Sunday and Olarewaju (2020) which focus particularly on the phonology of Nigerian children as they acquire their indigenous languages and/or the English language. These studies hYave advertently contributed to the literature on child language acquisition in NigerAia. R 1.2 Statement of the problem R The basic domains for describing child development have already beIeBn established and the under-investigation of the language domain, especially in Llow- and middle-income countries like Nigeria has been recognised (Fernald, et alN., 20 09; Wells and Stackhouse, 2016). The advocacy for growth monitoring by WHO, AUNICEF and other concerned bodies as a result of the high mortality rate of under-fivDe-year-old children in Africa has mainly been responded to by medical related researcAh and investigations (Akanbi and Anyarsor, 2014; Popoola and Adeoti, 2016; Eze, OBguonu, Ojinnaka and Ibe, 2017; Jimoh, Anyiam and Yakubu, 2018; Rossington, 1981) . I Thus, the language domain oOf chiFld development is still lagging (Wells and Stackhouse, 2016), despite the availab le phonological studies on Nigerian children’s language acquisition like AkpanY (2006), Ambrose (2010), Onwubiko (2011), Dopemu (2015), Adeniyi and AdeniIyiT (2017), Olarewaju (2018), Olarewaju and Sunday (2020), and Sunday and Olarewaju S(2020). This is because while these studies investigate children’s phonology, they have nRot done so from the perspective of describing children’s development. Hence, there iEs a dire need for a kind of research that will investigate children’s language aIcqVuisition with the purpose of describing their development. NSo far also, the available literature used in clinical phonological investigation of children is U based on British and American children’s language use (Rella 1989; Ball 2016; Bowen 2011), hence, there is a need to investigate and document Nigerian children’s early language use for this purpose. Therefore, this study investigates the language domain of child development through a focus on the phonological processes that accompany language acquisition within the first six years of life in one hundred and two Nigerian children. 3 Specific attention is paid to phonology because it is generally the first phase of language acquisition process in a child (Ball, 2016; Bowen, 2011). Prince and Smolensky’s (1993) Optimality Theory is adopted as theoretical framework. 1.3 Aim and objectives The aim of this study is to investigate the phonological processes of Nigerian childreYn’s English as means of tracing their language development and the objectives are: R 1. to identify the phonological processes in the children’s English language uAse; 2. to determine the constraints ranking for the emergence of these procesRses; and 3. to describe the children’s language development. IB 1.4 Scope of the study L This study examined the phonological processes of selAectedN Nigerian children in Lagos and Oyo States. Data were elicited from the children using an adopted text and through observation. Children between the ages of one andD six years were observed. A total of one hundred and two children, comprising boyBs anAd girls, were participants in the research. The following processes were teste dI: yod coalescence, assimilation, deletion, monophthongisation, vowel strengFthening, substitution, and gliding. 1.5 Significance of the s tOudy The study is especially Yimportant because it is in the area of phonology which has been identified as the mIoTst challenging for second language users and the most evasive to describe for linSguists. This study is an attempt to contribute to bringing structure and standard toR this seemingly evasive level of language, especially regarding children’s use. MoVre iEmportantly, this study contributes to child and developmental linguistics in Nigeria. ItI does this by providing important database to help describe the language of typically Ndeveloping children; this description can be used as a milestone for providing therapy to U children with speech delays. This can be done by ensuring that findings from this study are available to speech therapists that specialise in English as spoken by Nigerian children. More specifically, this study is very instrumental to the activities of rehabilitation centres, speech therapy clinics and concerned parents and guardians. This is because it presents information about the specific expected language progress of a Nigerian child. 4 Finally, this research is also significant because of its examination of phonological processes, a branch of suprasegmental phonology. Phonological processes are a further proof that language use is not a mere physiological activity but a psychological one which can only be adequately investigated using mental-oriented theories and tools of analysis. 1.6 Summary Y This first chapter of the study has presented an introduction to the research. It contRains the background to the study, the research problem, the aim and objectives, the scAope, and the significance of the study. The next chapter contains a review of conceptuRal and empirical literature, as well as a review of the theoretical framework. LI B DA N IB A OF TY RS I E IV UN 5 CHAPTER TWO LITERATURE REVIEW AND THEORETICAL FRAMEWORK Y 2.0 Preamble R This chapter focuses on the conceptual and empirical literature that provide releAvant context to the present study. These are language acquisition studies, pRsycholinguistics, phonological processes, Optimality Theory, germane empiriIcaBl studies and the International Phonetic Alphabet. L 2.1 Conceptual review AN 2.1.1 Language acquisition D Language acquisition is a process through whAich children develop the ability to understand and speak a language. It is a subcons ciIouBs process that requires no active effort from the child before acquisition is achievFed (Sureshkumar, 2002). Campbell and Wales (1970) describe language acquisitio nO as the process through which children achieve a confident control of their native laYnguage. This description however, does not capture all contexts of language acquisitionT, especially some contemporary second language environments of the English languaSge, Iwhere non-English children acquire English, or some variety of it, as their first lRanguage but do not have native-like competence. This coEntext nullifies Campbell and Wales’ (1970) definition because the authenticity of thIeV situation is undeniable in many second language environments. This contradictory Nsituation described in this definition engenders two important facts which are, first, it is not U only native languages that are acquired, rather, languages that are not culturally native or indigenous to a people can also be acquired by their offspring. Second, acquisition of a (first) language, especially one that is not native to the person, does not always feature native-like competence. 6 Furthermore, it is necessary to make a distinction between language acquisition and language learning. Language acquisition is different from language learning because the former is subconsciously carried out when children are still quite unaware while the latter is a conscious activity and is always a product of some form of formal learning as taught by someone or as self-taught (Kramiņa, 2000). Y These descriptions provoke two germane questions regarding English as spokenR as first language by some Nigerian children. Should it be considered to be a mother toAngue? Also, is it acquired or learnt? In response to these questions, insight is drawn froRm Kroon (2003) which proffers three intertwined definitions of mother tongue. TheI Bfirst is that a mother tongue is one’s native language which one speaks with one’s pLarents and/or guardians in the home. The second description is from the perspective oNf po litics and language policies where a language is deliberately chosen or created to sAuit the purposes of a mother tongue. Finally, Kroon (2003) describes mother tongue as Dthe official and standard language variety that is used as the language of instruction in sAchools. These descriptions give additional pe rIspBectives to the general and more widely known definition that a mother tongue Fis only one’s indigenous language. Therefore, these descriptions provide an answOer to the first question and that is the fact that English in Nigeria can qualify to be call ed a mother tongue among some families who use only English as the medium of coTmmYunication. This situation is not peculiar to Nigeria alone because Shaeffer (2020) alsIo notes that in Asia and some other regions, the younger generations speak a differeSnt language from their parents’ and grandparents’ generations, so that the language tRhe parents grew up speaking is different from the language that their children grew uEp with. V NInI response to the second question raised about whether Nigerian children acquire or learn U English as their first language, the answer to this question is not absolute and this is mainly because of the educational style that is generally upheld in Nigeria. From the very early age of two years, children are already enrolled in crèche and Montessori schools where they are exposed to some level of formal teaching and language learning. However, this cannot be said to account for all the language command that the children possess because English is 7 also spoken in their homes. Therefore, it appears that a good number of Nigerian children simultaneously learn and acquire English as their first language. Zaščerinska (2010) argues that foreign languages are always consciously learnt while native languages are subconsciously acquired, however, this definition seems too absolute as there are now second language contexts that make this description problematic. TYhus, the liberal and synonymous usage of both language acquisition and language learRning by scholars like Sureshkumar (2002) should not be acceptable. Therefore, in the cAourse of this research, language acquisition, which enjoys the focus of the research, isR used to refer to the subconscious process through which children imbibe the abIiliBty to comprehend a language and express themselves in it. L 2.1.2 Language acquisition studies AN Studies in child language acquisition began towarDds the end of the nineteenth century and it started with parents who kept diaries of thAeir children’s language development. These diaries were known as parental diariesI oBr “baby biographies”. More advanced research emerged as years went by and, based o n their peculiarities, these studies have been divided into three – diary, large sample anFd longitudinal studies. Matthews (1996) and Salim and Mehawesh (2014) provide sOimilar reports of this history and delineation. These are presented in a tabular foYrm below: IT RS E N I V U 8 Table 2.1 History of child language studies S/N Phase Matthews (1996) Salim and Mehawesh (2014) 1 Diary studies 1877 – 1930 1876 – 1926 Y 2 Large sample studies 1930 – 1957 1926 – 1957 AR 3 Longitudinal studies 1957 until now 1957 until now BR LI N(A uthor’s Conception, 2021) A BA D I OF TY RS I IV E UN 9 The delineation in Table 2.1 shows the progression of child language acquisition studies. First is the examination of a single child’s language development, whether by its parent(s) or a linguist; second is a study of a large group of children which provides more data to draw conclusions from. The third is a combination of the first two and it involves an observation of a much smaller group of children for a long period of time. The third mode of research provides a better opportunity for linguists to draw more robust and specYific conclusions on research objectives. Longitudinal studies are essential because theyR provide a better opportunity to explore the process of language acquisition – a procAess that has defied wholesome understanding for many years. A major part of theB studRies in the 1870s focused on the earliest stage of language development because it wasI thought that this stage has more information about the origin of communicative com pLetence (Matthews, 1996; Salim and Mehawesh, 2014). Findings are notably consisNtent across these investigations; one of these is that children’s first words are about thinAgs that fascinate them (Rice, 1989). There have been various attempts at explaining thDe procedure for language acquisition in children and these have yielded different BtheoAries. Some notable ones are discussed briefly in the subsequent paragraphs. I 2.1.3 Theories of languageO acqFuisition Some of the key theoYries within the discipline of child language acquisition are behaviourism, nativIiTsm, cognitivism, and socio-interactionism. BehaviourismS BehaviouriRsm is a movement in the field of psychology that was prevalent between the 1I92V0s Eand the 1950s (Field, 2004). John Watson and B.F. Skinner held the view that there Ncan only be speculations about the workings of the human brain, and that these speculations U can be based only on the external activities and behaviour that humans display. This movement holds strongly that human thought-processes depend largely on language and is a sub-vocal aspect of speech itself. It also perceives language acquisition as a set of habits which a child imitates from the parents and for which such child is rewarded. 10 Skinner, a prominent figure in the theory of behaviourism, proposes that acquiring a language has to do with linking an external stimulus to a person’s response. An instance of learning is the classical conditioning in which a person’s response is linked to a particular stimulus. Another learning strategy proposed by Skinner is the operant conditioning where rewards are used to reinforce or reward responses. Being convinced by this, Skinner (1957) asserts that children are able to acquire language by imitating adults’ utterances. HeYnce, parents provide the basis for imitation and encourage the children by correcting theRir errors and rewarding their correctness. RA Grammar is said to develop in the form of sentence fBrames into which words or phrases can be inserted. A proIcess of ‘chaining’ accounts for the way in which Lwords are organised in sequence, with the first word in the sentence providing a stimulus for the second, the secNond for the third and so on. This account considerably stretches what was originally understood by the DterAms ‘stimulus’ and ‘reinforcement’. BA (Skinner, 1957: p. 31) I Field (2004) also notes Skinner’s pFrov ision for some functions of the child’s language use. Imitations are labelled as echoic utterances, while mands are the child’s request for something. Tacts are the chi ldO’s response to a non-verbal nudge and intraverbal responses have no syntactic linTk toY verbal predecessors which engendered them. Chomsky (195S9) crIiticised Skinner’s hypothesis of imitation being a medium of children’s language aRcquisition. First, Chomsky counters that parents do not make the kind of mistakEes that children make and neither do parents correct the grammar of their children. HIeV notes that parents are more concerned about the correctness of the content of what the Nchildren say. Thus, behaviourism’s central idea that language is a set of acquired habits is U debunked by Chomsky. In contemporary research however, there is enough evidence to believe that imitation does have some role in a child’s language acquisition process (Fernández and Cairn, 2011); hence, Chomsky’s absolute refutation of the entire theory is considered premature. Also, Chomsky’s conclusion that parents do not correct the language of their children is rather untrue especially considering a second language context, where parents correct their children’s utterances. 11 Cognitivism Cognitivism perceives language acquisition as intertwined with general cognition and cognitive development (Field, 2004). As a child grows in years, this cognitive perception gradually becomes better and more improved. Piaget (1970) established this concept more when he stated that language comes about as a result of cognition and perception. Y In the same vein, Rice (1989) shows that cognitive and mental developmentR prompt language acquisition. However, she counters Piaget’s claim that cognition gArows before language does, stating that both occur roughly simultaneously. For instaRnce, toddlers are saying “all gone” at about the same time they are solving LageI-a Bdvanced tasks. She emphasises this when she says that “it is not the case tha t pre-linguistic conceptual knowledge always precedes and accounts for language Ndevelopment. Instead, the two domains seem to develop synchronously.” (p.151) A Despite both researchers’ convincing argumeAnts Dhowever, only part of their submissions are verifiable and that is the aspect about language expression, while the other part about cognitive growth and development o f IchBildren are largely subjective and subject to the physical but limited demonstratiFons of children. The internal workings and thought processes in the mind are s tiOll greatly undecipherable through theories and hypothesis, except through technoloYgical interventions. Socio-InteractioniIsmT Socio-InterRactiSonism views language acquisition as a joint effort from the child’s environEment, the child’s social instincts, pragmatic needs and relationship with the parents aIndV/or caregivers. Proponents of this approach agree that there is an innate aspect to Nlanguage acquisition but that this is not enough to explain the whole process. It is important U to note that both linguistic and non-linguistic interactions are a result of the child’s need to interact and communicate (Vygotsky, 1978). Further, socio-interactionists argue that stimulus is not as impoverished as Chomksy claims. They assert that adults’ self-correction, slow speech rate, increased intonation patterns, which are generally the features of motherese seem to help children understand 12 what is being communicated. This theory counters the assertion by nativism that children do not receive corrective feedback; the socio-interactionists put forward that corrective feedback is usually indirect – utterance recast, puzzlement, and grammatically correct responses. This theory balances out claims from some other approaches by acknowledging the meYntal processes that are committed to a child’s language acquisition process and also maiRntaining that the language spoken in the child’s environment is equally important. A gAood instance of this is children who, for different reasons, were shut out of any form of sRocial interaction in the early years of their lives and who could not understand or speaIkB any human language as a result. These children have the same mental capacity that othLer children have; however, because they were not exposed to any language, they are neither able to speak nor understand a language. This goes to prove that the languNage from the environment is as important as the mental ability of a child. DA Nativism A Nativism holds that language acquisit ioInB is genetically designed, that is, children are born with the innate ability to acquire laFnguage. Levelt (2013) lends support to nativism because of his conviction that since l aOnguage shares certain key similarities with other behaviours, then it must be controlled by something that is innate (Field, 2004). Central to the nativist theory is the notion YI Tthat children possess innate linguistic knowledge which makes them aware of the grSammatical principles in a language. This knowledge is combined with the linguistic dRata which they receive as input from the adults around to help them complete the acqEuisition process of language within five years. However, this stance has been faulted bIy Vmany researchers who counter “the universality of generative grammar, the autonomy Nof syntax in language processing, and the fundamental unlearnability of language” (Fernald U and Marchman, 2006: p. 1030). Crain and Lillo-Martin (1999) propose universal grammar as an alternative means to explain the acquisition of language by children. They make a distinction between linguistic performance and linguistic competence and base their theorisation on the latter which 13 focuses on the mental rules that determine language production, as against the former which is usually influenced by such speech factors like hesitation, distractions and the like. The timeless question of how children really come to master language within a few years is answered by these scholars in the simple claim that children are configured to do so. They report Chomsky’s earlier submission that the innate ability to master a languagYe is embedded in the child’s Language Acquisition Device (LAD); this enables themR to have an unconscious possession of the universal features of language as well as theA parameters and constraints of the language of their environment. Hence, as they acqRuire words from the input they receive, they map these words onto the existing princiIplBes they already have. This school of thought follows Chomsky’s (1996) submission. L Crain and Lillo-Martin (1999) propose a schema that sNhows the progression from the primary linguistic data (PLD) to the language acquisitiAon device and to the final state. This strict sequence shows the progression involved iDn language acquisition. First, a child is exposed to the language that adults speak aroAund it; from this language, a child is able to gather the input for the acquisition pro cIesBs. This input is called the PLD; it is subjected to the child’s innate ability for langFuage learning which is embedded in the LAD. This continues to progressively woOrk on the input the child has gathered until he attains the final stage which is the adult lang uage. Below is the schema: Inpu YIt T(PLD) → LAD → Final state (Crain and Lillo-Martin, 1999:5) The possibilityS of a child learning a language without being taught or instructed is expressed in the notioRn of universality of language (Crain and Lillo-Martin, 1995) and this assertion is insepEarable from Chomsky’s submission on the Language Acquisition Device which is mIoVre recently addressed within Universal Grammar (Chomsky, 1996). This principle holds Nthat typically developing children all across the world undergo the same stages of language U acquisition. However, the age at which these stages occur may vary by months across children. Therefore, it is safer for language development to be described in stages instead of in ages. Although these stages are universal, there are factors that make one child’s language acquisition process distinct from others. These factors are social, environmental, individual differences, and availability of PLD. Primary Linguistic Data could be directly 14 or indirectly presented through motherese or adult speech, explicit or implicit feedback may be provided, and there may be no feedback at all. Also, feedback may be either targeted at grammaticality or the truth-value of the child’s utterance. Despite these possible variations throughout the stages of children’s language acquisition, they are all able to subject PLD to LAD until they attain the final state. This typical achievement despite the peculiar colourations is what is referred to as universality of language (Crain and LilloR-MaYrtin, 1999). A Fernández and Cairn (2011) also designed a figure which is very similar tRo the schema by Crain and Lillo-Martin (1995). This figure contains the essential comIpBonents of the schema as well as the progression. Hence, the thematic content portrayedL in both illustrations is the same and this is that the input is subjected to the LAD whiNch in turn provides the output. DA A IBF Y O IT ER S NI V U 15 LAD Output: Input: (Language acquisition Device) Signals in the child’s Grammar Universal Grammar environment (speech, (principles , parameters) gestures) in one and ac quisition language or more than stra tegies one language Lexicon Y EXTERNAL INTERNAL AR Fig. 2.1 Schematic diagram of the relationship between external stimRuli and internal knowledge in language acquisition IB Fernández, E. M. and Cairns, H. S. 2011. Fundamentals of ps yLcholinguistics. New York: N John Wiley and Sons. AD A IB OF TY SI VE R UN I 16 Figure 2.1 shows the sequence of the processes involved in a child’s language acquisition. Although the innate biological attribute of every human is present in the child already, it cannot solely give the child a language, in other words, the child still needs the input – the language, before the LAD is engaged. The grammar that the child develops encapsulates such sub-areas like phonological components, syntactic and morphological components. These become more sophisticated as the child matures in age and experienceR in Ythe language. A Although children’s eventual language command is partly a product of tRhe environment, Crain and Lillo-Martin (1999) emphasise that the parents cannot be sIaBid to occupy a central position within this environmental factor. They cite an example oLf immigrant parents whose children master the language of the new environment, while th ey (parents) only struggle to attain the simple level of language command for basic purpNoses. Thus, with this, they prove that parents cannot be said to be the (only) languaDge mAodels for their children. Despite this limitation however, its impact on children shoAuld not be belittled. By extension, language developmen t IamBong atypically developing children who have varying degrees of difficulties withF language should also be described, not in terms of age but in terms of stages, stages Owhich, in their case, will be greatly varied by many months, even years, depending on th e severity of their challenge. Towards a justific YIatTion of the nativist theory of language acquisition, Crain and Lillo-Martin (1999) Scritique a series of mechanisms which have been previously submitted as basis for lRanguage acquisition. These are trial and error, corrective feedback, imitation, expansEion and motherese. 1I. V Trial and error: Crain and Lillo-Martin (1999) debunk the trial and error Nmechanism because of the universal process of child language acquisition. This universality U suggests that there is a system behind the process. In other words, if all children acquired language through a trial and error method, they would not share the same sequence. Also, the convergence on the same grammatical rules by children learning the same language is another proof that language is not acquired through trial. A further proof is that children make too few errors (in comparison with the totality of the possible errors in a language). 17 In concurrence with Crain and Lillo-Martin’s submission, children’s errors are quite systematic; hence, this nullifies the trial and error proposition. Apart from the criticisms presented by these authors, trial and error mechanism in itself presents a conscious and deliberate effort which is conspicuously absent in children at this stage. 2. Corrective feedback (negative evidence): There are objections to the notion Ythat children’s acquisition of language is as a result of the corrections they get from theirR parents and/or guardians. The research notes that parents are more concerned aboutA the truth in children’s language use than the grammar and that children do not alwayRs respond when adults attempt to correct their ungrammatical baby-talk. Fernández IanBd Cairn (2011) state that although some parents tend to think that they are obliged Lto correct their children’s ungrammatical expressions, majority of parents correct the content-accuracy of their children’s expressions rather than the structure. They alsoN observe that based on relevant research, correction by adults does not usually make mAuch of a change in the language of children. This is because the children acquireA a laDnguage on their own terms and in their natural phase and pace. IB However, there is reason to debaFte t his argument put forward by Fernández and Cairn (2011) because parents, especOially in English as a second language context, care about their children’s language as muc h as they do about the content. Therefore, parents, guardians and teachers in this cTlimYe correct their children and wards’ language use right from when these children are aIble to understand. Thus, it is logical to submit that the native language context in whSich these scholars have carried out their research prompts them to overgeneraRlise that this is what obtains in other language contexts. 3I. V EImitation: Another popular notion is that children develop their language skills Nthrough imitation of adults’ use of the same language. Crain and Lillo-Martin (1999) agree U that there may be some iota of confidence in this claim, however, they argue that there must be more to the process of acquisition than just imitation. They buttress this argument with children’s overregularisation of principles, culling an example from the addition of regular past tense morpheme (-ed) to irregular verbs, hence producing words like “goed” or “runned”. They further note that this overregularisation continues in the children’s language for months, sometimes years. In addition, Fernández and Cairn (2011) explain 18 that the extent of utterance imitation varies from child to child and that even those who imitate do not do it all of the time neither do they acquire language more quickly than those who do not imitate. 4. Expansion: When children use their short and simple forms of sentences, their parents and guardians often extend these into full-fledged sentences – this is caYlled expansion; it may also be subsumed under motherese (Rice, 1989) but has been seRparated here to accommodate the ongoing review of the critique carried out by CraiAn and Lillo- Martin (1999). Expansion cannot be a core mechanism for language acquRisition and there is an experimental research to prove this (Crain and Lillo-Martin, 199I9B). In this experiment, there were two groups of children, one group had adults expandLing their utterances, while the adults in the other group did not expand the children’s uNtter ances. After the study, there was no improvement in the language of the experimAental group relative to the control group. It is thus concluded that parents’ expansioDn of their children’s utterances does not contribute to the children’s language developmAent in any significant way. This experiment records that that the dIifBference recorded was insignificant, however, the term insignificant is really a relatiFve and subjective term especially because no statistical analysis was provided. Also, iOt is already a fact that no single one of these mechanisms has a monotony of the language acquisition process, thus, it is necessary to record the impact, no matter how minimTal,Y that each mechanism has on the entire process. 5. MotherSeseI: This refers to adults’ simplification of their language use so as to come down to thRe comprehension level of children. Its features include an emphasis on objects and peEople within their vicinity, that is, concrete reference to here and now, a limited vocVabulary, a lot of paraphrasing and frequent repetitions, simple and well-formed NseIntences, and a slow speech rate with pauses between utterances and after content words. U Other features include clearer articulation, higher pitch, use of full names instead of pronouns, and gestures to attract attention (Rice, 1989; Crain and Lillo-Martin, 1999; Duncan, 2013). Motherese also involves semantic contingency – a situation when the child mentions an item and the adult expands on it; this helps the children to expand their vocabulary. These features of motherese are not all universal, however, when used within the construct of a particular culture, it reflects how infants and young children are 19 addressed. Crain and Lillo-Martin (1999) submit that previous studies showed no significant difference between the use or otherwise of motherese with children. They add that simplifying language structure and its use only makes acquisition difficult for children. Fernández and Cairn (2011) corroborate this claim by stating that although infant-directed exaggerated prosody may help children distinguish the phonemes, words and phrasal boundaries of their language easily, without this assistance, they would have acquired thYese eventually. Also, the children spoken to in motherese do not acquire language fasRter than those who are not. Therefore, motherese cannot be the sole explanation for hAow children acquire language. BRI Although Fernández and Cairn (2011) establish that these discu sLsed mechanisms (trial and error, corrective feedback, imitation , expansion, and motherese) are not very central to a child’s language acquisition, they emphasise that input shNould not only be available, but should also be interactive. To corroborate this point, Athey cite Sachs, Bard and Johnson’s (1981) study of two children of deaf parents – JimD and Glen. These children were unable to adequately acquire spoken language BbecaAuse their parents were incapacitated in this regard, and the only exposure to langu agIe they had was with the television which was non- interactive and therefore inadequaFte. However, after series of oral interaction with them, the language of the older chi ldO, Glen, got extremely better. This then raises the TqueYstion of what it really means for language use with children to be interactive, if not pIartly to expand their language use, involve the use of motherese, listen as children imSitate, and correct them when they need it. These are major activities that culminate Rinto the interaction that Fernández and Cairn’s (2011) advocate for. In other words,E Fernández and Cairn (2011) have under-acknowledged the impact of these mIeVchanisms in children’s language acquisition process. Since they have established the Nimportance of an interactive Primary Linguistic Data, their critique of the above U mechanisms seems myopic because these are what help the available linguistic data to be interactive. Hence, the combination of expansion, motherese, corrective feedback and imitation are what really make language to be interactive for children’s acquisition process. Therefore, it is safe to say that no single mechanism here is responsible for acquisition like Crain and 20 Lillo-Martin present; rather, it is an incorporation of them all. Therefore, this realisation faults the schema and diagrammatic illustrations that both studies have put forward because such factors like interaction and cognition are missing. Hence, there is a need for a revision of these presentations and the current study presents this below: Y AR IB R L N AD A IB OF TY SI VE R UN I 21 Input Interphase Output Oral language LAD Adult Language Gestures/Signals (Language Acquisition Device) Y Sign language Target LanguaRge COGNITION A INTERACTION (Imitation, motherese, R corrective feedback, expansion) LI B Fig. 2.2 Factor interactions in language acquisition AN (Author’s Conception, 2021) AD IB OF TY RS I E NI V U 22 Figure 2.2 has three phases which are the input, interphase and output. The input phase presents the language tokens that a child is exposed to from the very beginning of a language acquisition process. These language tokens could be oral or gestural, and they are made available when adults and older children use this language around them or when they watch the television and are exposed to some social media platform. This input phase does not capture written language because acquiring the ability to read written language invoYlves a conscious effort from the child, thus, this skill is not captured in the figure aAboveR. After a child is exposed to these language expressions, three different coRmponents come together in the interphase. These are cognition, interaction and LAIDB. Cognition involves their mental processes, interaction has to do with how the peoplLe around them engage this language with them, and the innate language acquisition devi ce is the naturally endowed ability to acquire a language. All these are jointly manoeuNvred to achieve the acquisition process. DA This figure incorporates LAD, cognition Aand interaction in the interphase because Fernández and Cairn’s (2011) “exte rnIaBl” and “internal stimuli” were removed. This removal was necessitated by the dFifficulty in classifying “interaction” as either external or internal stimuli. FurthermoreO, the interaction factor was also expanded to accommodate imitation, motherese, corre ctive feedback and expansion which the previous schemas ignored. This accomTmoYdation is permitted because all the mechanisms contribute their quota to the overaIll achievement of acquisition. It also is noteworthy that although the varying theoreStical approaches to language acquisition view the subject from different angles, theRy all agree on the fact that a child’s cognitive domain works on an ambient languaEge to achieve language acquisition. V NTIhe current study maintains a stance that is similar to that of socio-interactionism. It U combines the mechanisms under the interaction and subsumes it within the interphase, along with cognition and LAD. This proposed theory is called a cogno-interactionist theory of language acquisition. This approach is so called because it draws on cognition, LAD and interaction as the core elements that describe the process of child language acquisition. The under-listed are the adopted tenets of this proposed approach: 23 1. A child’s cognition is important to language acquisition and, as it grows, the cognition increases and so does comprehension and expression of language. 2. A child’s language acquisition is propelled by its need to respond and interact with the society; hence, both linguistic and non-linguistic interactions stimulate language acquisition. 3. Continued exposure to interaction is important for a continued and progresYsive acquisition of language. R 4. Children, especially before the age of five, are naturally endowed toA understand language without being exactly taught. RIB None of the above listed items supersedes the others; each of theLm exists in dependence to the others and possesses equal importance. Children like Jim a nd Glen in Sachs, Bard and Johnson’s (1981) study are proof that only having theA innNate ability to learn a language is not enough for language acquisition. Also, atypically developing children who suffer mental and/or physiological inabilities are prooDf that having a language being spoken around a child is not enough to acquire tBhe lAanguage either. All the factors have to work together for a child to acquire a langua gIe adequately. 2.1.4 Logical problem of laOnguFage acquisition (LPLA) This problem simply reYfers to the question of how children are able to master language rapidly despite the IlimTited experience they are exposed to in the early phase of life. In other words, it asks Sthe fundamental question of how children are able to avoid the numerous potential uRngrammaticalities in language use (Rice, 1989; Crain and Lillo-Martin, 1999; HyamsE and Orfitelli, 2015). Logical problem of language acquisition has to do with how cIhiVldren come to know the inappropriate and improbable structures in their languages even Nwhen they are only exposed to the possible structures by the adults. How do they know U what not to say and what is and is not correct? To answer this germane question, Crain and Lillo-Martin (1999) rely heavily on the principles of Universal Grammar. Rice (1989) describes what Baker (1979) pointed out as three observations of this basic question: 1. there is no negative evidence available to children; 24 2. children tend to overgeneralize linguistic rules and make mistakes similar to the grammatical sentences…; and 3. finally, the constraints in a language are quite arbitrary and not readily predictable and thus seem not to be easily learnable p. 150 Rice’s (1989) and, Hyams and Orfitelli’s (2015) reference to this question shows that Yit is still as relevant in the 2000s as it was in 1980s. Rice (1989) further reveals thRat since children are unable to answer questions despite the manifestation of theiAr expressive language skill within the first year, data for research has to be gatheredB fromR their utterances and what is said to them. In other words, researchers have to garneIr what they can out of what the children say since they cannot really directly answer q uLestions when asked. Another question that LPLA asks is whether or not childreNn know how correct or incorrect their utterances are. The present answer to this queAstion is negative. Children are not consciously aware to the extent of deliberately pDroducing correct or incorrect structures. However, what has been established is that thAey are able to acquire words from what they hear around them and model th eIir Bstructures after them. This explains their overgeneralisations of certain ruFles. In answer to the logical problem of language acquisition, therefore, it is su bOmitted that the children do not really avoid ungrammaticality, rather, they speak it as Ythey hear, process it and arrive at ungrammaticalities which are modelled after the grTammatical structures they hear. 2.1.5 The dilSemmIa of the “how” Rice (1E989R) states that the real research problem is finding out the process of how children extVract from the language they hear and perceive the conventional linguistic rules of their nIative languages. The answers of cognitive and social underpinnings or innate linguistic Ndevices only show the phases; they really do not answer the question of how. As of 1989 U when her article was written, Rice says scholars have focused their energy on describing the children’s linguistic abilities and patterns of acquisition but not really on how the language is acquired; three decades after, this question is still being asked and is yet to be adequately answered (Guolaugsdottir, 2016). Rice (1989) states that there is a clear answer to the question of how cognitive mechanisms help children resolve the learnability problem 25 or arrive at language-specific knowledge. She faults the linguistic models that support an innate linguistic learning device as not having satisfactorily specified how such device would work and how it interfaces with the rich network of general cognitive mechanisms available to even very young children. Resolution of these problems constitutes the most challenging of the current questions about language development. Y Lust (2006) reiterates that “[l]anguage acquisition is an inherently intellectual feaRt in that children do complex theory construction. The growth of language is mediated iAn the human species by complex symbolic computation” (p. 1). Stork and WiddRowson’s (1974) innateness hypothesis corroborates this, and draws the following geIneBralisations: a. All human beings learn a language L b. All human languages are equally learnable c. All human languages are different on thNe surface but all have certain underlying features Ain common, which account for their being “learnable” d. These features which are uAnivDersal to all human beings are the key to what is innate From the review, it is evident that child IlaBnguage acquisition is still a research-worthy field as there are still questions that arFe in need of answers, especially regarding the specific activities that children undergo to acquire language by just listening. It has been established that there is a universal fra mOY ework that makes acquiring language possible for children (Chomsky, 1996; Crain and Lillo-Martin, 1999; Fernández and Cairn, 2011). However, the question remains –I wThat is the nature and structure of this universal framework? S Locke’s (2R009) study aptly provides a path to an answer for the salient question of how which Eis raised by Rice (1989), Lust (2007) and Guolaugsdottir (2016). Locke avers that uInrVavelling this unknown from the angle of genetic studies has also been futile because Ngenes do not work in isolation but require external signal which in this case is language. U He argues the need for linguistics to situate language within an evolutionary and developmental framework to birth a new branch of linguistics – evolutionary development linguistics which can then merge the resources of language studies with that of evolutional development so that the exact process a child uses to acquire language can be unravelled. This suggestion is modelled after the merger of evolutionary biology and developmental biology in order to arrive at a richer understanding of the modifications of development as 26 well as its processes and production of new features. In other words, Locke (2009) calls for a joint effort between linguistics and evolutional development. He says: A new branch of linguistics, “Evolutionary Developmental Linguistics” (EDL), would be able to import findings from research in the other fields of cognitive and developmental neuroscience, and export theory and data to those fields. The Y result is almost certain to be a better understanding of language itself. R (p. 36R) A He further refers to Chomsky’s description of language acquisition aIsB a biological process. Therefore, if linguists who have made great headway in child laLnguage studies are able to combine their efforts with scholars in science fields, greatNer results would be achieved. In line with this, Matthews (1996) also expressed optimiAsm about the future and potential of neural network models in the analysis of languDage. Therefore, Locke (2009) calls for linguists to exploit maximally, the potential inA Evolutionary Developmental Linguistics so that scholars outside linguistics do not eInBd up postulating theories in language studies. As helpful as this suggestion is, thFe si tuation really is that there are already collaborations between linguistics and some Odisciplines of science like psychology, neurology, computer science, hence such arYeas like psycholinguistics, neurolinguistics and computational linguistics (Kager, 20T03). He adds that it is the objective of linguistic disciplines to expound how the humanS mIind processes linguistic knowledge and that this endeavour is not the direct respoRnsibility of a formal theory. This opinion corroborates Locke’s submission that linguisEtics needs to partner with other disciplines to achieve a more comprehensive undVerstanding of language acquisition. N IU 2.1.6 Psycholinguistics There are quite a number of scholars who have tried to trace the history of psycholinguistics. Garnham, Garrod and Sanford (2006) state that psycholinguistics was recognised and established as a discipline around the end of the nineteenth century and that this development helped medicine record phenomenal growth. A seminal example of this 27 development is Broca’s and Wernicke’s original descriptions of aphasia which happened during this time. Also, Coleman (2013) traces the inception of psycholinguistics to the 1950s when an interdisciplinary seminar was organised in 1953. Levelt (2013) submits that the history of psycholinguistics can be traced back to 1762. Psycholinguistics is an interdisciplinary exploration of language using the resourceYs of psychology and linguistics. It examines the connection between linguistic behaviRour and psychological processes, and the process through which language is acquiredA (Fernández and Cairns, 2011). Field (2006) asserts that this branch helps to unraveRl the connection between the human mind and language. It does this on an indivIiBdual basis (although individuals share mental properties), then these studies are brouLght together in an attempt to figure out the correlating features, and how the human mind works. Hence, the norm is established primarily for the purpose of knowledge aAnd tNo carry out more studies (Field, 2006). D Tanenhaus (1989) notes that psycholinguistiAcs investigates three unique areas which are how language is acquired, how it isI uBnderstood and how it is produced. Language acquisition is the focus of developFmental psycholinguistics; language comprehension and production constitute the corOe of what is done within experimental psycholinguistics. Levelt (2013) also corrobo rates this classification of psycholinguistics into three sub- classes. However, FiTeld Y(2006) identifies six areas which are central to psycholinguistics; these are languagIe processing, language storage and access, comprehension theory, language and Sthe brain, language in exceptional circumstances and first language acquisitionR. Although Field’s (2006) classification appears to be more diverse, it still does not capEture certain areas that may be considered central to psycholinguistics, for instance seIcVond language acquisition. Also, some of the classifications he makes could have been Nmerged into the broader ones provided by Tanenhaus (1989) and Levelt (2013). U The area that is central to the current study is first language acquisition. This area of psycholinguistics attempts to solve the age-long puzzle of the entirety of stages that children undergo to eventually acquire their first language. Fernández and Cairns (2011) corroborate this by stating that psycholinguistics is geared towards comprehending the acquisition of language, its use and comprehension by different people and its 28 representation and processing in the human brain. They state that there is a direct link between psycholinguistics and some other disciplines like developmental psychology, cognitive psychology, neurolinguistics and speech science. Psycholinguistics draws heavily from theories in cognitive psychology but is somewhat limited in its use of experimental techniques unlike studies in second language acquisition (Field, 2006). Y Levelt (2013) writes that evolutionary biologists, including Charles Darwin, theorRised the phylogenetic development of speech in humans as well as the potentiaAl of innate recapitulation in ontogenesis. Simply, these scientists propose that languaRge development in humans is in phases and stages. The psychologists perceived the oIntBogenesis of language as a sign that coginitive development was taking place; hencLe, a child’s command of language was a phase of its developmental psychology. Over th e course of various research in the area of psychology, it became central to find oAut hNow a child’s expanding bank of sounds gets connected to pictures and abstract concepts. How does the child’s word bank grow; how does it master the formation of wordsD, their meanings and how they are used? These various inquisitions into child languageA acquisition worked differently until the late nineteenth century. However, while t heIyB worked differently, they had a common feature which was that many of these invFestigations were carried out by parents – diary studies, and this form of study still e xiOsts till now (Levelt, 2013). Levelt’s account of theY history of psycholinguistics reveals the tussle for theoretical supremacy betweenI Tempiricism and nativism for the explanation of how a child acquires language. HowSever, he notes that there is more agreement than disagreement on this issue. The first arRea of agreement among the different researchers is the order in which children acquireE sounds. It is usually preceded by the ‘outer speech’ and followed by the ‘inner spIeVech’. By the outer speech is meant the sounds which carry no semantic burden, while Nby the inner speech is the sounds which are combined to form words. The sequence of the U activities between these phases is general among children, however, the timing differs greatly from child to child. 2.1.7 Stages of language acquisition 29 Language acquisition is in phases and each of the phases shares similarities with an adult’s language use. Also, these phases are similar across languages and Chomsky (1959) shows that the principles applicable in the consecutive stages of these languages are universal. However, Rahimpour (2004) argues that this universality is only a possibility, not a certainty. This probability is evidenced in motherese which although is present in many languages, manifests differently in various cultures. The timing of language developmYent in children may vary by months but, the events and stages themselves are generaRl. Thus, the sequence of activities in the acquisition process is unmistakably consistenAt (Crain and Lillo-Martin 1999; Fernández and Cairn, 2011). This extensive simIilBarity R in the stages of children’s language acquisition is a pointer to the fact that language has a biological predisposition to it, and that humans possess an innate skill wh icLh brings about the process of language acquisition (Fernández and Cairns, 2011). N In earlier studies of developmental psycholinguistDics, NAelson (1977) revealed that attention was not paid to the sounds and gestures that a baby made until the first sentence was uttered. This was so because it was perceived that the Asentence was the basic and only fundamental medium of communication. However , IthBis claim has since been debunked as it has been discovered that babies express thFemselves and communicate even with their pre-verbal sounds. Before a child begOins to produce the actual language, it first begins with interpretation and compYrehe nsion. In fact, a child exhibits comprehension earlier than it expresses itself (CrIaiTn and Lillo-Martin, 1999; Levelt, 2013; Salim and Mehawesh, 2014). For comprehenSsion at the early stages, a child depends heavily on extra-linguistic cues to understandR. In their second year of life, how quickly a child understands language use is a centralE indication of how well they will use the language. Nelson (1977) notes that “child laInVguage has its roots in the social communication system and in the individual cognitive Nsystem and both must be coordinated and matched to the language system of the U community” (p. 571). The mystery behind how a child acquires the syntax of their language is demystified when he states that a child uses the same learning strategies employed in other areas. The pertinent question here is: is it the same with phonology acquisition? However, as similar 30 as children’s language acquisition processes are, their idiosyncrasies interfere and influence the outcome of the process (Nelson, 1977). Children acquire sounds that are found in all languages first, afterwards, they learn those that are particular to their own languages (Szczegielniak, 1982; Crain and Lillo-Martin, 1999). Vowels are learnt and produced before consonants. Perhaps, this is the case becaYuse vowel phonemes are more easily produced since the major articulator needed is theR tongue while the consonants need more articulators than just the tongue. AccordinAg to Nelson (1977), the consonant acquisition is also usually ordered: for manner Rof articulation – “nasals, glides, stops, liquids, fricatives, affricates” while for pIlaBce of articulation – “labials, velars, alveolars, and palatals” (p. 573). Szczegielniak L(1982) corroborates more specifically that at this stage, children perceive more sounds t han they produce and when they have some difficulty producing a particular sound, thNey replace it with another. The children do not carry out these replacements haphazaArdly, rather, they are rule-governed, although unconscious. However, these rules cannoDt be said to be universal. Also, consonant clusters are simplified, syllables are duplBicateAd and final consonant phonemes are elided. Although, O’Grady and Choo’s (20 0I1) findings show that stops occur before other consonant sounds in manner of Farticulation, they agree that regarding the place of articulation of consonant phonOemes, labials are produced first. However, it is important to be cautious about being Yrigid concerning the order in which these consonants occur among children, because fiIrsTt, the languages of the world do not all have the same set of phonemes, and second, chiSldren will acquire the easiest phonemes from the whole lot that they perceive around theRm. Rice (1E989) reports that among children, the early social functions of language show reImVarkable similarity but these functions are not universal. Although they produce words Naround the same time, and the same sort of meanings are encoded in early words and U sentences, there are still variations at the individual level, in terms of rate and manner of acquisition, combination of words, choice of only content words or not and so forth. All this prove that a child’s idiosyncrasy comes to fore when acquiring a language. From the beginning, children engage language as a social tool to seek attention, protest, greet, request, refute and so on. Hence, it is evident that they are socially and intellectually 31 motivated to acquire a language for their pre-linguistic communicative needs. The social setting validates the effectiveness of a child as a communicator – whether this communication is carried out with words or just sounds (cries), it is perceived and treated as meaningful. Rahimpour (2004) identifies two important skills used to explain the developmenYt of language in children; these are the receptive skill and the expressive skill. The rReceptive skill often comes before the expressive skill because as soon as a child is bornA and begins to hear different noises and/or sounds, its receptive skill is already at workR. The expressive skill is put to use when they begin to produce sounds for communicIatBion. Fernández and Cairn (2011) break down language developme nLt in children to their ages and each of these classifications is discussed below. N 1 Before birth DA According to research (Hepper and ShahiduAllah, 1994; Graven and Browne, 2008), the hearing ability begins to develop in chIildBren from the second trimester at about eighteen weeks of gestation period and by the outset of the third trimester, the foetus is able to respond to auditory stimulus. OFernFández and Cairn (2011) corroborate this assertion with a reference to a study carried out by Barbara Kisilevsky and her colleagues in 2003. It was discovered that a foeTtus Yprefers its mother’s voice over that of a stranger and it shows this through an increaseI in the rate of heartbeat and body movements. This occurs at full-term at about thirty-Seight weeks of gestation. 2 EFirRst few months AIfVter an approximate six to eight weeks, infants begin to distinguish human sounds, Nespecially the mother’s, from others. They show recognition with a smile and engage their U expressive skill which begins with response to stimuli like hunger, pain, pleasure, satisfaction, discomfort and so forth, but more importantly however, this is when they begin to produce sounds at will through babbling, crying, and cooing. Babbling is the production of linguistic sounds but which lack any linguistic meaning (Stork and Widdowson, 1974). 32 Duncan (2013) describes their sounds at this stage as reflexive because it is a reflection of their internal state. During this time too, they recognise human faces and distinguish voiced from voiceless consonants – a distinction called categorical perception (Crain and Lillo- Martin, 1999). In addition, researchers attribute babies’ recognition of faces to their innate abilities. However, the question of whether babies’ innate abilities is the sole explanation for their recognition of human faces needs to be asked because logically, experience hYas a contributing factor to this feat too. A child is able to recognise this face because Rit sees it every day over the weeks; therefore, it becomes familiar with the face and recAognition sets in after a while. Potentially, the same can be said of a child and a pupBpy. ROver the months of seeing a puppy, a child becomes so used to that puppy to thLe poIint of recognition and being at ease around that particular puppy. Therefore, it seems unmerited and incorrect to attribute all the credit to innateness hypothesis only. N 3 Approximately six to ten months DA In the sixth month, infants make a wide raAnge of sounds during playtime (Stork and Widdowson, 1974). There is a repetiItioBn of open monosyllables, and sometimes, the common phonemes are replaced with other less common ones. Duncan (2013) calls this reduplicative babbling while SOtorkF and Widdowson (1974) call it the reduplication stage. Even deaf children alsoY dis play this feature at this stage, although they do not get any primary linguistic IdaTta from the environment (Crain and Lillo-Martin 1999). Therefore, this shows thatS this activity begins as “an internally driven behaviour, not a response to external stRimulation” (p. 26). The universality of the process of language acquisition is furtherE emphasised here because children in different communities undergo the same aIctVivities and produce roughly the same sounds until about ten months when the Nenvironmental factors begin to affect their sounds so that they become peculiar to the U immediate environment. This activity is what Duncan (2013) describes as non-reduplicative babbling. They also babble using the stress and intonation contours in their immediate environment. At this stage, deaf children stop babbling. By the ninth month, the infant responds to simple words, gestures and movements. 33 Regarding the non-linguistic development, the child laughs and wriggles on the floor till they begin to crawl and sit up (Crain and Lillo-Martin, 1999). According to Duncan (2013) however, laughing and vocal play begin at about the fourth month, but this cannot be really countered because, as Brown (1973) states, these developments vary by months from child to child. Y 4 Approximately one year R By the twelfth month, children respond to commands. They engage their expAressive skill when they actively participate in language and speech (Stork and WiddowRson, 1974). They begin to produce their first words at about this time; these words areI uBsually in reference to entities in their immediate environment, for example, mummy , Lmama, daddy, dada, baba, biscuit, poopoo and so forth (Crain and Lillo-Martin 1N999; Duncan 2013). A major percentage of the first fifty words of a child’s vocabuAlary will be nouns and will include names of people, objects, animals, foods, body paDrts and so on (Duncan, 2013). Alongside this is the development of jargons – incompreAhensible words which babies say. During this stage, the holophrastic language is en gIagBed – this refers to single-word utterances (Stork and Widdowson, 1974). F Gestures are also introduced , Ofor example, stretching out for a cup of water when thirsty; these gestures may alsoY be accompanied with words. The physical development includes cruising, standing, IwTaving and so on (Crain and Lillo-Martin, 1999). Children acquire an average of nine words per day. Evidently, they are able to do this without specific tutoring. They absorRb thSese words from the conversations and utterances around them; while at it, they alEso gather some level of word meanings, formation and grammar. 5I V Approximately one and a half years UNAt around eighteen months, a child already has a vocabulary of about twenty words and jargons. Also, Saffran, Senghas and Trueswell (2001) aver that at around this age, there is usually a spurt of new words and word combinations by a child and they understand about one hundred and fifty words. The intonation pattern used for these words is a single pattern – falling tunes on each word and with spaces in-between them. They seem to isolate lexical items from a string of sentences in the primary linguistic data and their vocabulary grows. 34 Soon, the two-word sentences are put in an intonational group with a fall tune at the end. The physical development includes walking, self-feeding, scribbling lines and colouring. 6 Approximately two years (Stage I) By the twenty-fourth month, a child is able to combine words into sentences, and go from producing simple sentences to complex ones (Stork and Widdowson, 1974). Brown (19Y73) proposes that development should be grouped into stages as from the age of two, sRo that a child who is approximately two years is said to be in Stage I. He notes thaRt thAe length of a child’s utterance correlates very much with the stages of development. An average of the length of the child’s spontaneous utterances is used to determineI Bthe Mean Length of Utterances (MLU). This is not tantamount to the child’s gram mLatical competence; rather, it gives a clue about the child’s language development relatNive to other children. Children’s word order is usually patterned after that of adults andA at this stage, there are more content words than grammatical ones. Physical develoDpment includes activities like running, pulling off socks, scribbling in circles and so Aforth (Crain and Lillo-Martin, 1999). 7 Approximately two and a ha lfI toB three years (Stage II) Mean length of utterances (MLU) iFs two point two-five (2.25) words with an approximately nine-hundred-word (900) v oOcabulary. The child uses more function words and also includes progressiveT andY past tense forms of verbs in their utterances. Other features at this stage include overIregularisaton, reference to absent entities and past events, inquisition about the envSironment and so forth. At this stage, the child is yet to fully master grammaticRal morphemes, this will happen over the course of several stages. Physical develoEpment includes jumping, removing clothes, drawing closed figures and so forth (CIrVain and Lillo-Martin, 1999). UN 8 Approximately three to three and a half years (Stage III) The child’s MLU increases to two point seven-five (2.75) words and their vocabulary increases to one thousand and two hundred (1,200). More function words are amassed at this stage and the child uses syntactic transformations which enable it to form questions 35 from declarative statements, although, mistakes are still inevitable. Brightstone (2000) also submits that by this time, children are able to use language quite elaborately. The physical development at this time includes riding a tricycle, washing their faces, hopping on one foot and so on (Crain and Lillo-Martin, 1999). 9 Approximately three and a half to four years (Stage IV) Y MLU and vocabulary increase to three point five (3.5) words and one thousand Rand five hundred (1,500) words respectively. They produce multi-clause sentences Alike relative clauses, complement clauses, conjoined clauses, however, overreguIlBarisa Rtion still occurs. The physical development cues include catching a ball with hands, using a pair of scissors, doing the jigsaw puzzle, telling stories and so forth (Crain and Lillo-Martin, 1999). 10 Approximately four to five years (Stage 5) AN At this stage, the average length of the child’s wDords, that is the MLU, increases to four point zero (4.0) while the vocabulary accommAodates up to one thousand and nine hundred (1,900) words. The child learns and IusBes more grammatical words like conjunctions, subordinate clauses, temporal words li ke “before”, “after”, “when” are also used. The child now engages its peers in coOnverFsations and displays metalinguistic abilities like self-correction and defining wo rds. This shows a conscious awareness of the properties of language on the partT of Ythe child. Their display of metalinguistic skills marks a milestone in their developmeInt unlike before the present stage when they could not pinpoint or tell what was wroSng with their constructions. The child’s non-linguistic and physical developmeRntal cues include activities like learning basic mathematical notions (Crain and Lillo-MEartin, 1999). V N IU 11 After five years From the age of five years, children’s sentences become more complex; between five and ten years, vocabulary increases more slowly. Also, they learn exceptions and stop overregularising (Crain and Lillo-Martin, 1999). 36 An examination of Stork and Widdowson’s (1974), Crain and Lillo-Martin’s (1999) and Duncan’s (2013) exposition of the stages of a child’s language development shows that there are many grounds of convergences. Hence, it can be easily concluded that the universality of language acquisition indeed cuts across different climes, cultures and languages. However, the heavy reliance on the principle of innateness to explain the specific processes of how children acquire language leaves a lot of knowledge outsideY the purview of linguistics, thus, proving more difficult to use linguistic theories to expRlain the intricacies of language acquisition. RA On the one hand, the studies mentioned above counter Lust’s (200I6B) submission that by their third year, children would have acquired their first languag eL. Hendriks and Spenader’s (2006) study on the other hand, corroborates the developmental classifications above when she states that language acquisition begins with the pAre-laNnguage phase which is between zero and twelve months, and that within this period, there are different phases of acquisition like the egocentric babble phase (4-6 months) andD the social babble phase (7-12 months). After the pre-language phase is the early lanAguage phase which covers the twelfth to the thirtieth month. She further states tha t IthBe sound production stages move from crying, to cooing (first month), to babbling (sFixth month), to production of first words (tenth – twelfth month), overextension and ovOerrestriction. However, as elaborate as her discussion of the stages is, it does not covYer th e period when children begin to incorporate the syntax of their languages. IT The abilities thSat put language development in motion are naturally endowed to a child and are shapedR by the environment but the degree to which the environment participates in this processE is the bone of contention among scholars and their research (Hyams and Orfitelli, 2I01V5). Saffran, Senghas and Trueswell (2001) create a metaphor to explain how children Ncomprehend and engage linguistic resources to create their own language. Children have to U carve out words out of the acoustic flow of utterances that adults release, words that are only seldom surrounded by pauses. They have to do this before these words can be assigned to any object in the physical world. Despite the tangible difficulty in this process, these scholars report that from the age of seven months, children are able to do it. While answering the question of how children separate utterances into words, Saffran, Senghas 37 and Trueswell’s (2001) experiment shows that the children were able to do this when they were presented with “miniature artificial languages that embody specific aspects of natural language structure”. They submit that once an infant has been familiarized with a sample of this language, a new sample, or a sample from a different language, is presented to the infant. Subtle measures of Y surprise (e.g., duration of looking toward the new sounds) R are then used to assess whether the infant perceives the new sample as more of the same, or something different. In this A fashion, we can ask what the infant extracted from thRe artificial language, which can lead to insights regardIinBg the learning mechanisms underlying the earliest stages of language acquisition. (1) L Thus, infants detect when a new language has been introdNuced and this is manifested with a moment of surprise. According to them, infants aDre abAle to recognise words by identifying the syllables. Further studies from these researchers show that they used the eye movements of children to detect children’s comprehensionA of a given instruction. However, as elevated as this study is, it does not seem total lyI rBeliable as some of the analyses and findings are vague. Also, it is not clear how thFe claims in the results have been verified, for instance, the number of children who wOere used in the study is absent. In addition, it is uncertain whether this number is rYepresentative enough for making generalisations. Brightstone (2000)I aTvers that by the time children are three years old, they can use language quite elaborateSly, adding that the process may seem simple and ordinary, but an in-depth examinatioRn shows otherwise. A typical child’s ability to comprehend and use language is limitedE mainly to their mental and linguistic development. Locke (2009) corroborates this bIy Vsaying that an effective language acquisition is not the function of the language or the Nspeakers around a child, rather, it is a function of the child and the influence of their genes. U Rice (1989:13) emphasises this by saying that in “normally developing children, language emerges spontaneously as a means of talking about what they know so that they can accomplish social goals important to them.” One of the remarkable facts expressed by Brightstone (2000) is also conveyed by Rice (1989) and this is that in their second year of 38 life, children begin to talk quite coherently, and they become adept at using language to express themselves and little or no apparent teaching is needed to achieve this. 2.1.8 Phonological processes Phonological processes are the changes that occur in sounds when they follow one another in close succession in an utterance. Sounds in isolation are stressed, however, when tYhey occur together in words and with other sounds, the sounds flow into one another, Rbringing about phonetic and phonemic adjustments like deletion, coalescence, reductioAn, insertion, assimilation, aspiration, nasalisations, labialisation and so forth (Pike, 19R48; Cruttenden, 2001). These changes are what are referred to as phonological pIroBcesses or connected speech processes. Although these two terms refer to the saLme phenomenon, it was discovered during the course of this research that phonoloNgical processes are often used to refer to the changes that occur in children’s utterancAes as found in studies like Bowen (2011) and Ball (2016), while connected speech Dprocesses are used in reference to adult utterances as found in research like Nolan aAnd Kerswill (1990), Roach and Widdowson (2001), Oladipupo (2014) and so on. T hIeBrefore, the term phonological processes has been adopted to refer to the changes thaFt occur when sounds co-occur in children’s utterances. Phonological processes cons tiOtute a major part of language development in children; this is because children, as a reYsult of their ongoing development, do not have complete control of some of their artiTculators such as the lips, tongue, teeth, and jaw, hence, they end up simplifying cerStainI words and phonemes for ease. These simplifications occur in different ways and RAkpan (2004) sectionalises them into three groups which are substitution, assimilEatory and syllable structure processes, while Roach and Widdowson (2001) group thIeVm into assimilation, coarticulation and elision. However, the shortcoming with these two Ngroupings is that there are certain phonological processes that do not fit into any of these U groups, while there are some others which fit into more than one group. Thus, this breeds a problem of classification. These processes occur among both typically and atypically developing children as well as adults, although their instantiations are often different. Ball (2016: p. 14) describes phonological processes as a 39 mental operation that applies in speech to substitute for a class of sounds or sound sequences presenting a common difficulty to the speech capacity of the individual, an alternative class identical but lacking the difficult property. Levelt (2013) describes the situation as when difficult sounds are simplified or replaced by less difficult ones. Sounds in citation forms or in isolation are often differenYt in pronunciation compared to when they occur in the company of other sounds. GRiegerich (1992: p. 249) notes that “citation forms are a form of speech that is, to sAay the least, somewhat idealised”. R Wundt (1900) provides the result of what happens when soundIsB occur together and influence one another. He describes the major ones as assimil atLion and dissimilation; the former refers to how sounds become more similar as agaiNnst when they occur alone. The latter is the opposite, sounds become less similar thanA they would be when they occur in isolation. D Also, about English phonological processes,A Sutomo (2012) states specifically that they occur both at the word and the phras alI lBevels and that these processes are caused by the influence of syntactic factors. AtF the word level, morphemes are combined with other morphemes, bringing abou t Osome form of change on at least one of the morphemes. Concerning these processes, Kerswill (1985) avers that consonant and vowel sounds are transformed or even TdeleYted sometimes; sounds which occur beside others are modified to sound like the oSnesI in their environment. These are all processes that occur when there is a progressiveR sequence of words in speech (Nolan and Kerswill, 1990). Oladipupo (2014) adds thEat phonological processes are a manifestation of intrinsic human tendencies to react toI Vthe challenges of speech by simplifying difficult sounds. Some of these processes are Ndiscussed in the subsequent paragraphs. U 1. Vowel strengthening: This process affects English central vowels which are usually weakly produced. Vowel strengthening occurs when this otherwise weak vowel segments are produced with more amplitude and acoustic energy to strengthen them. Thus, central vowels are produced like front or back vowels (Bowen, 2011). 40 2. Substitution: Substitution refers to total replacement of one sound with another one. There are several varying motivations for the occurrence of this process. In clear terms however, it involves a replacement of one sound with a totally different one (Ball, 2016). 3. Deletion: Deletion is also known as elision and it can affect both vowel (apocope) and consonant phonemes (syncope) in different parts of a syllable. Thus, thereY are different types like final consonant deletion, initial consonant deletion. ADeleRtion can also affect whole syllables, especially weak ones. Often, the sylRlable containing schwa is elided in the child’s speech (Bowen, 2011). 4. BMonophthongisation: This process refers to the reductioIn of a diphthong or a triphthong to a pure vowel. This means that one or two c oLnstituent vowel phonemes in a diphthong or a triphthong, respectively, is reNmoved in order to achieve this process. In the case of a diphthong, the two vAowel segments may also merge into one to produce this pure vowel. The end Dproduct of this process is that only one vowel phoneme emerges (Bowen, 201A1). 5. Assimilation: Assimilation is thIe Bmodification of a sound in order to make it similar to some other sound in its env ironment. This modification can occur in the area of voice, place of articulationF or manner of articulation. A major class of words in the English language tha tO are affected by assimilation are the allomorphs for regular plurality morTpheYme and regular past tense markers (Laver, 1968). There iSs alIso a directionality involved in assimilation process, that is, it can be forwRard or backward in direction (Zsiga, 2011). When assimilation occurs in Esegment 2 because of segment 1, then assimilation is progressive, however, when IV assimilation occurs in segment 1 because of segment 2, then it is said to be N regressive. Assimilation can also be bidirectional but this is very rare in the English U language and its varieties. There are different types of assimilation and these get their differentiation and nomenclatures from the type of sounds that are involved in the assimilation. Some of them are palatalisation, labialisation, and nasalisation (Chapitré, 2005). Labialisation, patalisation and nasalisation are when non-labial, non-palatal and non-nasal sounds, respectively, are produced as such because of their proximity to these sounds. 41 Zsiga (2003: p. 1934-5) says about assimilation processes specifically, that they are innate because “the phonetic motivation for such processes is clear, and the motivation works in the direction of making speaking easier”. She goes on to expound the reasons assimilation occurs – this is because articulatory organs cannot move quickly enough to prevent from ‘running’ into one another, therefore, the activity of one cannot be ended quickly enough for another to begin, hence, thYese organs form a continuum of activities, which results into a continuum Aof soRunds. Apart from voice, laryngeal features of obstruents and sonorantsR also sometimes interact. Zsiga (2011) states that local assimilation is thIe Bcommonest type of phonological process, and this means that it is cenLtral to the workings of phonological theories and phonological features.N Ass imilation also occurs with vowels when they occur in the context of nasaAl sounds; this is called nasalisation. While in the case of continuants, stops uDsually transform into continuants when preceded by continuants. A stop can Aalso change to a fricative; this is known as spirantisation and can be perceiIveBd as a form of assimilation of the feature. Zsiga opines that the assimilation o f place of articulation is easily the most common phonological process, espeFcially the nasal place of assimilation. 6. Dissimilation: This pOrocess involves the enhancement of the distinctions between sounds so thTat sYounds become more auditorily different from the others in the environmenIt. This process enhances perception and is found in a limited set of words iSn the English language. A good instance is when the morpheme -al is added to sRome nouns. This closed set of words is a residue from Latin (Chapitré, 2005). 7. VEGliding: This usually involves sliding from the production of one segment into I another one. An articulator moves from one point of articulation to another, hence N producing a similar but still different sound. This process occurs when /u/ becomes U /w/ and /i/ becomes /j/ (Bowen, 2011). 8. Yod Coalescence: Coalescence is when two sounds merge into one. This newly produced sound then shares phonological and/or phonetic qualities with the two original sounds responsible for its emergence. Weisser (2005) notes that yod coalescence has to do with the blend of an alveolar plosive segment or a fricative 42 with a back vowel. Weisser (2005) makes a distinction between yod coalescence and u-coalescence, however, this separation appears unnecessary because both sub- classes involve the same sets of sounds for coalescence to occur. 9. Voicing: Voicing is said to have occurred when an otherwise voiceless consonant phoneme is produced with the vibration of the glottis because of its proximitYy to another voiced segment. Since vowel segments are always voiced, this procRess only affects voiceless consonant segments (Bowen, 2011). A 10. Devoicing: When a segment that is ordinarily voiced loses its voiceR feature because of its proximity to a voiceless segment, the process that IhBas occurred there is devoicing (Bowen, 2011). L 11. Stopping: When air from the lungs is not released Ngrad ually as found in a fricative but released suddenly and explosively like a Aplosive, then stopping has occurred (Bowen, 2011). 12. Reduplication: This refers to the repetiDtion of part of a syllable or the whole syllable (Ball, 2016). A 13. Cluster reduction: Cluster re duIcBtion occurs when the consonants that make up a cluster are reduced, thus siFmplifying the cluster (Ball, 2016). 14. Epenthesis: When a Osound is inserted in-between two consonants, it is called epenthesis. It is Yoften used to break up consonant clusters although this is not the only contexIt Tin which it occurs (Bowen, 2011). 2.2 EmpiriScal literature on child language studies EmpiriEcal Rresearch on child language acquisition is reviewed in this section to shed more ligVht on what has been done so far on the subject matter. Prescriptive and descriptive studies oIn child language acquisition have been reviewed in subsequent subsections. Also, the third UNsubsection is dedicated to research on phonological processes. Finally, the fourth subsection contains an appraisal of all the reviews. 2.2.1 Prescriptive studies on child language Alerechi (2011) studied consonant substitution in the language of an Ikwere child. The study submits that children’s simplification of the language being acquired can contribute 43 to so much language divergence that adults will then need to learn children’s language. However, this assertion is not entirely valid because it is gathered from empirical and conceptual review that the language of children undergoes phases; it is constantly subject to change as their anatomy, especially their speech organs, develop. The peculiar simplification of sounds that they engage in is only for a while, because they assimilate towards adult speech as they grow older. For these reasons, this study’s conclusion Ythat children’s language use can lead to divergence cannot be accepted. Hence, there isR never a need to learn children’s language because children eventually attain adult targAet language, except in the case of atypically developing children. IB R Five to eight-year-old bilingual children of Igbo and Yoruba lanLguages were the focus of Onwubiko’s (2011) language development studies. The stud y concludes that the Igbo- Yoruba bilingual children have articulation problems whNich manifest phonologically as omission, substitution, consonant deletion, and epenAthesis in their use of the English language. The instances presented of these proAcesDses show that they are found even among adult speakers too. Hence, it is erroneouBs to label them as articulation problems because children will always eventually acquir e Itheir ambient language in a matter of months. Hutauruku (2015) studied theO lanFguage acquisition process of children between one and three years of age. It adopt s an eclectic theoretical approach, drawing on insights from proponents of different Ytheories some of whom are Chomsky, Lyons, Bolinger, Fromkin, and Steinberg. ThisI tTheoretical combination made the analysis to lack a clear focus because of the proponeSnts’ divergent views on children’s language acquisition. Furthermore, the study termRed the participants’ phonological processes as problems and errors caused by lazinesEs. On the contrary however, it is incorrect to label these features as problems or eIrrVors because these are actually features of progress in the process of language acquisition N(Rice, 1989; Crain and Lillo-Martin, 1999; Lust 2006; Fernández and Cairns, 2011). A U closer examination of the instances cited in the work showed that the children were following the existing pattern in that language, although these patterns were sometimes over-obeyed. Hutauruku (2015) focused on the children’s language use from the perspective of errors and problems, and not from that of development. 44 Etim, Dada, and Bassey (2018) tested fifteen children for their use of five phonological processes. Their utterances were recorded and fed into the PRAAT Language software for analysis. Simple percentage and generative phonology were also used for analysis which shows that the children exhibited the five tested phonological processes. However, the study also concluded that the children were deficit in the pronunciation of affricates, velar frontings, liquid gliding, final consonants and consonant clusters, and that the proceYsses were impairments. The research further suggested that remedying intervention stRrategies and therapy sessions be instigated for the children. The study’s RconAclusion and recommendation prove that many linguists and researchers peBrceive phonological processes in children’s language use as errors that one should be wIary of. However, they are actually ways through which a child’s language domain and , Lby extension, their general development can be described. N The phonological features of ten Lagos-resident NigeriAan children of educated parents were examined in Ademola-Adeoye (2019). The idDentified phonological processes were substitution, voicing, cluster reduction, deletAion and reduplication. The study concluded that the children were more disposed tIo Bacquire the Nigerian variety of English than the Received Pronunciation (RP). HoFwever, this conclusion is rather unexpected; children acquire their ambient langua gOe and so are unlikely to acquire the RP when it is not spoken around them. Thus, cYomparing the children’s phonological processes to Received Pronunciation showIsT that the study is prescriptive. Its failure to consider the contemporary fluidity of the SRP (Jowitt, 2016) already faulted the study’s objective. Apart from this, it was erroneRous to expect the children to articulate the RP even when they had never been exposeEd to it. This is because children acquire the variety of language that is spoken around theVm. NIThe following sub-section contains descriptive studies on phonological processes and U children’s language acquisition. These studies cover both monolingual and bilingual children. 2.2.2 Descriptive studies on child language 45 Akpan (2006) studied the assimilation process of an Ibibio monolingual child and the following assimilations were discovered – homorganic nasal assimilation for both vowel and consonant phonemes, nasalisation of vowel and consonant sounds, voicing, labialisation and substitution. The classification of substitution as an assimilation process raises questions. Also, the study only showed the frequency of the instances, the peculiarity of the phonological processes was not adequately discussed in relation to the IbYibio monolingual, even when more generalisations and deductions could have been drawRn from the data. Akpan (2006) concluded that assimilatory processes should not existA beyond five years, but if it does, such child should be made to undergo therapy bBecauRse it is a sign of disorder/impairment. However, this is not entirely the case because pIhonological processes are not features of child language only, they are also found iLn the language of adults, although they manifest differently. Hence, it is importantN to always add this clause so as not to get parents, guardians and teachers unnecessarilAy worried or mislead researchers and language enthusiasts. D Also, Ambrose (2010) investigated one GwonAg child’s use of semantic overextensions and phonological processes. Gwong is one oIf Bthe languages used in Kaduna State, Nigeria. The phonological processes studied aFre assimilation, reduplication, substitution, consonant cluster and syllable simplifi cOation. Some of this study’s findings are that the participant used more nouns, followYed by verbs. This corroborates what has been established in the literature that childIrTen acquire more nouns before other word classes because these are usually the names of the people around them, the food they like to eat, their favourite pets and toys anRd sSo forth. Hence, it is easier for them to acquire and use these words before other wEord classes. This is buttressed by some of the child’s first words: “Hyedah (her elder siIsVter), uncle (in reference to the researcher), ball, car, drink, see, take, finish, it, me, yes” Nand “hey”. All these confirm that language acquisition principles are universal among U typically developing children. The study is however extremely limited by its sample size, thus, many of its findings are not generalisable. Menn and Vihman’s (2011) study on the features of child phonology leans towards an emergentist model of feature acquisition. Data for the study was collected from about fifty children who spoke Estonian, Finnish, Japanese, Welsh and some Germanic and Romance 46 languages. The study deliberately countered the innateness hypothesis. In answer to its question as to whether children’s phonological features are innate, the study proposed that the combination of the language that children hear, their cognition which is still developing, and their articulatory organs which will also keep developing for a while, is responsible for their language production. It refutes the claim of nativism that an innate feature is solely responsible for language acquisition. Therefore, the study concludes that languYage acquisition features are inherent only in the sense that they are biological procesRses, not arbitrary ones. They are also products of an interaction between the different Acomponents of the system described above. The study clearly states that the acquisiBtionR of these features is not “pre-experiential cognitive given” (p. 284). LI Eren (2015) examined the acquisition of the alveolar tap (/ɾ/) in sixty three-to-five-year-old children in Gazientrop province, Turkey. At age five, the Nchildren were yet to attain 75% of /ɾ/ occurrences in words; also, four-year-olds did nAot produce up to 50% of the total of the /ɾ/ occurrences in words. The study’s objectivDe was an attempt to get normative data towards a standardisation of Turkish articulatiAon. However, the realisation of this objective may be very minimal because many oIf Bchildren’s phonological features are temporary. They have not yet completely maFstered how to manoeuvre their speech articulators, but when they eventually do, th eiOr utterances will change too and become more similar to an adult variety. Therefore,Y it becomes a futile effort to describe their language for the purpose of standardisation IbeTcause this phase is not a permanent one. Furthermore, the study is silent about whether the data collection from the three different parts of the province has any effect on thSe children’s production of the alveolar tap /ɾ/. At 5:0-5:11 years, the children were yEet toR fully acquire /ɾ/, showing that the children’s language was still developing at aIgeV 5. Having established in its literature that the acquisition of /ɾ/ is usually completed by Nage 6 among Australian, English, Spanish, Arabic and American-speaking children, the U researcher ought to have included six-year-olds in the sample size. Thus, the study’s submission that the children were experiencing some kind of delay is incorrect. Phonological processes in a non-standard variety of Portuguese as produced by two- to six- year-old children were the focus of Quelroga, Rosal, Solva and Cordeiro (2015). The utterances of two hundred and two kindergarten children who were enrolled in public 47 schools in Recife were recorded, transcribed, coded and tabulated for easier descriptive analysis. The results of the analysis showed that the most common phonological processes were consonant cluster simplification, syllable reduction and simplification of final consonants. Liquid consonant simplification, frontalisation of palatal and posteriorisation of palatal processes ceased before the age the researchers projected. However, the participants continued to simplify consonant clusters and the researchers related this toY the effect of socio-cultural issues in the society. The study suggested that sociolinRguistics should be considered when examining children’s phonology to prevenRt miAsjudgement. Notably, the study recorded progress in the children’s phonology, buItB it is silent about how long it gathered data from the participants; this could have been instrumental in replicating a similar research elsewhere. L Yousofi and Ashtarian’s (2015) study identified three staNges of language use from birth; these are the protolanguage, prespeech and true worAd stage (adult language). The study explored the percentage of proto and true wordAs inD the Kurdish of a child between 9 months and 36 months of age. They discovered thBat a major percentage of the child’s proto words were produced in the first six months oIf the child’s second year; however, the articulation of these words continued into the tFwenty-ninth month during which the child also used true words very frequently. The stuOdy infers that factors that influence a child’s lexicon use are both general and languaYge- specific; also, these factors are strongly related to the people around the child as Twell as its immediate needs. As a child advances, more variation is observed in itsS lanIguage use. Although the study is a longitudinal one, the distinction of child languRage into three parts (protolanguage, prespeech and true word) is not really shown in the dEata provided, rather, only two aspects were focused on which are the protoword and theV true word. NIIbrahim and Ibrahim (2016) investigated morphological and phonological processes in the U language acquisition process of six Hausa children. The different types of reduplication found in the data include reduplication of a CV syllable, harmonising consonants, word reduplication and imitation of onomatopoeic sounds. However, the study was neither explicit about its methodology nor in-depth in its discussion. The study was silent about the 48 ages of the children, thus, it is impossible to tell whether the production of only a CV syllable structure is age-appropriate or not. Adeniyi and Adeniyi’s (2017) study described the consonant sequence reduction of four children who were simultaneously acquiring English and Yoruba. Of these four children, two were acquiring English primarily and Yoruba secondarily (group 1) while the oYther two had Yoruba as their primarily acquired language (group 2). It was discovered thRat group 1 deleted the most sonorous consonant in the sequences, while group 2 reduceAd consonant sequences through coalescence. These findings are precise and point to tRhe need to carry out a larger investigation on children’s phonology because, the limiIteBd sample size means that the findings are not generalisable and could have been inf luLenced by the participants’ idiosyncrasies. N Olarewaju (2018) explored English syllable acquisitioAn in ten bilingual children of English and Yoruba languages between the ages of 1;7 anDd 5;0. A fairly regular pattern of syllable structure acquisition was discovered amonAg the children despite variations in their acquisition rate, physiological maturationB and exposure to new words. Although the study states that the interference of Yoruba Iin the children’s acquisition is minimal, it is not explicit about this. This is OespeFcially important because of the different phonotactic structures of both languages. The subjects also simplified new and complex words into their marked structures througYh deletion, epenthesis and substitution. Although these processes are not the focus ofI thTe study, they introduce a new dimension to how children’s acquisition processes can bSe investigated especially because these processes are instinctive. The effect Rof sonority on consonant cluster production among bilingual children of Yoruba andV EnEglish was the focus of Olarewaju and Sunday (2020). The participants were found toI prefer less sonorous obstruents to the highly sonorant consonants whenever a consonant Ncluster occurred. To achieve this preference, the children deleted some segments, and these U deletions occurred more at the onset where single consonants were more preferred. However, the study’s sample instrument dwelt more on onsets than on codas because there were five and two words testing onsets and codas, respectively. Thus, the conclusion that the participants deleted more consonants from onsets could be as a result of this imbalance in instrument. 49 Sunday and Olarewaju’s (2020) investigation of the acquisition of complex coda among bilingual children of English and Yoruba showed a kind of progression in the articulation of children from 1 to 5 years, so that children who were three years and below still showed a great variation from that of adults, while those who were between four and five years had immensely assimilated towards the ambient language. However, the study’s treatment of the productions of these two quite different groups using the same scale seems erroneYous and problematic because both groups actually exhibit different phases of lRanguage acquisition. Specifically, children who were three years and less reduced theAir consonant clusters greatly, while children who were four and five years did noIt. B R 2.2.3 Phonological processes L Zsiga (2011) focused on local assimilation as found in Ndifferent languages around the world; the study gives an exposition on the intricAacies of local assimilation, clearly distinguishing it from long-distance assimilation. DThe study culled relevant instances from different languages like English, Russian, YorAuba, Turkish, Japanese, Ancient Greek, Zulu, and a myriad of other languages. The IinBstantiation from languages as many as this only proves further that phonological proce sses are features of language that can be found in all the languages of the world. OF Epenthesis is another veYry c ommon phonological process and Hall (2011) describes it as the insertion of a voIwTel in an utterance to adjust a word whose structure does not match the structural provision in a language. This researcher notes that there is no known reason for the phonotRacticS structures that engender epenthesis but adds that sometimes, a particular orderinEg of consonant phonemes prompt the intervention of an epenthesis no matter what thIeV syllable requirements of that language is. She cites Côté (2000) who argues that Nepenthesis occurs because of the need to make some consonants much more audible than U they are. However, audibility is not the only motivation for epenthesis; sometimes, the phonotactic provision of a language does not permit consonant clusters, thus, a speaker resorts to epenthesis in order to obey this phonotactic rule. A good instance of this is the case of consonant cluster prohibition in Yoruba, and its permission in English. 50 Sutomo (2012:11) employs the generative phonology theory to explain the phonological process of assimilation and other morpho-phonemic changes in English language. The study concluded that “one rule can apply in many other niches of the same phonological context”, hence, “language is [a] really rule-governed behaviour”. These applications help to understand the phonological changes that happen when sounds co-occur, however, Oladipupo (2014) opposes this idea when he argues that “phonological processes Yare phonetically motivated and not rule-governed as generative phonology proposesR” (p. 2). He also states that “these processes are results of innate human tendencies to reAspond to the difficulties of speech by simplifying difficult sounds”. With these, SutBomoR’s application of generative phonology to phonological processes seems inadequaLte. I Oladipupo (2014) studied aspects of sound modifications Nand simplifications in speech in order to discover features of connected speech in StanAdard Nigerian English. To this end, he applied natural phonology as theoretical framewDork and engaged three hundred and sixty educated Nigerian English speakers as sampAle population. He identified three groups of connected speech processes and labelBled them dominant, minor and idiosyncratic processes, where only the dominant pIrocesses were recognised as typical of Nigerian English speakers and were thereforFe referred to as the standard. Oladipupo faults the earlier works of phonological pe rsOpective on Nigerian English as only carrying out an identification of these pYhonological processes without any attempt at determining their extent of use. TherIefTore, he set out to identify, quantity and categorise these processes in Nigerian EngliSsh in order to decipher a standard. However, the study was silent about the specific lanRguage groups that were sampled as well as their distribution. JosViah Eand Soneye (2014) set out to identify some phonological processes in the spoken laInguage of some one hundred university students from five federal government-owned Nuniversities in the country. The aim of this research was to add to existing literature, U research and data on the educated variety of the Nigerian English. Their inference that educated Nigerian English (NE) is largely endonormative and indicative of an emerging Nigerian English variety only further confirms what has already been established in the literature on the subject matter. This study is inherently prescriptive, rather than descriptive, in its comparison of NE to standard British English. The method of analysis is also 51 questionable only words were identified, not specific sounds or processes as expected. The study concludes that assimilation is “a normal occurrence in Nigerian English and in most second language situations” because of some factors as mother tongue interference, slower tempo of utterance, idiosyncratic patterns of realisation, influence from socio-cultural and educational standards, social status, speech style, among others. The case however, is that assimilation is a normal feature in connected speech even in a native speaker’s languYage use and such use does not have to be as a result of any of the reasons that thAe stRudy has listed. R Utulu (2014) focused on the monophthongisation and vowel lenIgtBhening processes in educated Urhobo English using the paradigms of moraic theor yL as anchor for the analysis of the study. The study set out to find out the motivation Nfor the speakers’ lengthening of the first vowels of standard British English closing dipAhthongs (/eɪ/ and /əʊ/). The analyses revealed that the reasons for this lengthening arDe to preserve the weight of the deleted closing vowels and to reflect the components Aof the falling fundamental frequency contour of English final open syllables. TBhe discovery of this deletion within the monophthongisation process subs taIntiates Oladipupo’s (2014) suggestion that phonological processes are attemFpts at simplifying sounds that prove to be difficult in pronunciation for speakers. U tOulu (2014) goes further to generalise that the two motivations underline the simplificaYtion process of the same set of closing diphthongs in many non- native English accenTts of Africa and Asia. However, the study which was carried out on a minority tribe iSn NIigeria does not have enough premise to make such generalisation about non-nativeR English accents in Asia and the entirety of Africa. 2.2V.4 EAppraisal of literature review NTIhese reviews show the extent and depth of research carried out within the purview of child U language studies and phonological processes, outside and especially within Nigeria. They create the needed contextual background within which to situate the current study. Some of these studies (Alerechi, 2011; Hutauruku, 2015; Onwubiko, 2011; Ademola-Adeoye, 2019; Etim, Dada and Bassey (2018) are prescriptive because they perceive children’s phonological processes as problematic, erroneous and adulterated, and stipulate that these processes should be either like that of adults or like Received Pronunciation. 52 However, studies like Akpan (2006), Ambrose (2010), Onwubiko (2011), Adeniyi and Adeniyi (2016), Ibrahim and Ibrahim (2016), Olarewaju (2018), Olarewaju and Sunday (2020) and Sunday and Olarewaju (2020) explored the phonology and phonological processes of Nigerian children, hence contributing to child language studies in Nigeria. These studies described what occurred in the language of the participants, thus contributing to the literature of what a Nigerian child’s phonology is like. Eren (2015), Quelroga, RoYsal, Solva and Cordeiro (2015), and Yousofi and Ashtarian (2015) have also provided lRiterature for the expansion of child language studies. RA These studies are far from exhaustive of what can be done regardiIngB Nigerian children’s English phonological processes. In fact, some of these stud iLes investigated children’s indigenous language use (Ambrose, 2010; Ibrahim and IbNrahim, 2016) while some others studied bilingual children (Adeniyi and Adeniyi, 2016, Olarewaji, 2018; Olarewaju and Sunday, 2020; Sunday and Olarewaju, 2020). Also, Asome were longitudinal (Quelroga, Rosal, Solva and Cordeiro, 2015; Salim and MeDhawesh, 2014) while others were cross- sectional. These studies, despite their wealth Aof contribution to child language studies, are lacking in a deliberate effort of using thIeBse language nuances to establish or describe the development of a child, particularlFy a typical Nigerian child. Also, whereas, both methods are beneficial to any languag eO study, these studies used either to sample their participants. Therefore, as a means Yof tracing children’s development, this study set out to explore phonological proceIsTses of Nigerian children’s spoken English, using cross-sectional and longitudinal daSta collection methods. 2.3 IntRernational Phonetic Alphabet TIhVe In Eternational Phonetic Alphabet was developed sometime in the nineteenth century to Nprovide a uniform pronunciation system for the languages of the world. This is done by the U assignment of a symbol to each phoneme of languages. This Alphabet proffers a uniform representation of the spoken forms of the world languages. It has been revised occasionally by the International Phonetic Association (IPA). The alphabet can show both phonemic and phonetic transcriptions (Augustyn, et al. 2015). The International Phonetic Alphabet is presented below: 53 Y AR R LI B DA N IB A O F Y SI T R VE UN I 54 Y RA R IB L N Fig. 2.3 The International Phonetic Alphabet A Augustyn, A., Bauer, P., Duignan, B., Eldridge, AD. Gregersen, E., Luebering, J.E., McKenna, A., Petruzzello, M., Rafferty, J.P., ARay, M., Rogers, K., Tikkanen, A., Wallenfeldt, J., Zeidan A., and Zelazk oI, AB. 2015. Britannica Encyclopaedia. Chicago: Encyclopaedia Britannica. 404. F 2.4 Theoretical framew oOrk: Optimality Theory (OT) Optimality Theory ThasY been adopted as the theoretical framework to explain the instantiation of proIcesses which the participants have adopted. Therefore, the following paragraphs are Sdedicated to discussions on, reviews and the use of OT. OptimaElityR Theory is a linguistic theory proposed by Alan Prince and Paul Smolensky (1I9V93). The theory was originally designed to answer questions in phonology but is now Nalso applied to other linguistic areas like syntax, semantics, sociolinguistics, and historical U linguistics (McCarthy, 2007). In essence therefore, OT is a broad theory that covers the different aspects of grammar (Kager, 2001). Optimality Theory is an offshoot of Generative Grammar, a theory whose emphasis is on the description of the structure of language and universal principles; this description is done using experimental observation of linguistic typology and how language is learned, whether 55 children’s first language or adults’ second or third language. Although OT came out from generative grammar, it is fundamentally and essentially dissimilar from it in more ways than one. First, OT is based on the assumption that universal constraints are violable while previous models worked based on the premise of parametric variation of inviolate principles (Kager, 2001). Y This theory thrives on constraints rather than rules like other approaches to phonRological analysis. Generally, OT deals in the different operations and transformations thAat take place between the insertion of an input into a grammar and the eventual productRion of an output. These inputs are evaluated based on the hierarchy of constraints in tIhaBt language; this is to say that the same set of constraints are available to all languagesL, however, they are ranked differently in each, hence their difference. These lead to a ch oice of one structure which violates the least number of highly ranked constraints.A N BA D I OF TY I ER S V Candidate 1 I Candidate 2 Input GEN EVAL OUTPUT N Candidate 3 U Candidate 4 Fig. 2.4 A schematic grammar at work 56 McCarthy, J. J. 2007. What is optimality theory? Linguistics Department Faculty Publication Series. Massachusetts: University of Massachusetts Amherst. p. 3 Y ARR LI B N AD A IB OF TY SI VE R UN I 57 Certain components are central to the workings of OT and these are GEN, CON and EVAL. GEN generates candidates which are evaluated by EVAL based on the ranking of constraints (CON) in that language (McCarthy, 2007). In other words, EVAL determines which candidate emerges as the optimal candidate. Figure 2.4 illustrates these relationships. GEN generates an infinite list of candidates based on what it receives in the input – this is described as the richness of the base. This set of candidates is subjected to evaluation whYere candidates who do not meet up with the requirements of the constraints, esApeciRally, the highly ranked constraints are screened out. EVAL uses the hierarchy of constraints in a language as the basis for determining which candidate is harmonic (McBCarRthy, 2007). Thus, all the components of OT are universal except for the different rIankings of constraints which distinguish one language from the others. L Linguistic structures can be described as having two featureNs based on markedness, and this is whether they are marked or unmarked. A linguisticA structure is said to be marked when it has a different and unique feature which stanDds it out from the other structures in a language. “Unmarked values are cross-linguiAstically preferred and basic in all grammars while marked values are cross linguis ticIaBlly avoided and used by grammars only to create contrast” (Kager, 2001: p. 2). F Furthermore, a structure’s mOarked or unmarked state is not usually randomly decided, rather, it is a functioTn oYf the different but interwoven systems in a language. The major components of OTI are discussed further in the subsequent paragraphs. 2.4.1 GERN(eSrator) This coEmponent, on getting an input, generates an infinite set of all possible realisations; thIeVse possible outputs are called candidates. There is a basic and important assumption that Nthere are no language-specific limitations on the generated candidates as long as the U elements of composition are licit and within the language – this assumption is referred to as the richness of the base. For instance, when dealing with prosodic structure, the expected licit elements include mora, syllable, foot, prosodic word; when dealing with morphology, such licit elements include root, stem, word, affix etc; when dealing with syntax, licit elements include x-bar structure, heads/complements, and specifiers (Kager, 2001). 58 Different possible structures are permitted to be generated here because they will still be subjected to assessment according to the constraints in the language under consideration. GEN has a feature known as freedom of analysis and this affords it the opportunity to generate candidates freely and even repeatedly. Since GEN is considered to be a universal feature, it is important to lend it the freedom it requires to generate all the possible options for the input in a language. The grammar of a language involves how the constraintsY are ranked, and this ranking in turn determines which candidate will be ruled asA theR optimal candidate by EVAL (Kager, 2001). R 2.4.2 CON(straint) IB After GEN has generated the candidates, they are subjected to Lthe constraints as they are ranked in the particular language. Constraints are the criteNria that an output must obey for it to be considered a bon-a-fide part of a language. ThAis requirement can either be obeyed or violated by the candidate (Kager, 2001: p. D9). These constraints are found in all languages, that is, they are universal, howeverA, they are ranked differently in all languages, hence, the difference in the languages oIf Bthe world. CON helps to decide which candidate should be the harmonic candidate. No netheless, it is important to note that the violation of a constraint is not equal to ungrammFaticality, neither is the satisfaction of a constraint equal to grammaticality. Rather, coOnstraints are inherently competitive and so are always in conflict so that an oTutpYut will always have to violate some constraints but satisfy some others for it to emeIrge as an optimal or a harmonic candidate. The conflicts among these constraints are Susually resolved by ranking them (McCarthy, 2007). A moreE higRhly ranked constraint is satisfied above another constraint which is not as highly raInVked; this proves that the idea of a well-formed structure is relative to different languages. NWhat is well-formed in language A may be ill-formed in language B; it is all a function of U which constraint is ranked higher or not in one language or the other. In simple terms, constraint ranking is the reason that languages are different from one another (Kager, 2001). There are two types of constraints – the faithfulness constraint and the markedness constraint. The faithfulness constraint mandates that the optimal candidate must match the input; in other words, the surface structure (output) must tally with the underlying one 59 (input) in a particular way. In other words, it prevents the harmonic emergence of structures that do not resemble the input. Faithfulness constraint is described as conservative because of its requirement for a look-alike between the input and the output. Examples of faithfulness constraints include the following: 1. The output must preserve all segments present in the Y input. 2. The output must preserve the linear order of segments in R the input. A 3. Output segments must have counterparts in the input. 4. Output segments and input segments must share vBalueRs for [voice]. (Kager, 2L001I: p. 10) Markedness constraint on the other hand imposes a particNular well-formedness pattern on the output, such that changes are motivated and impressed on the output, making it different from the input. This class of constraints is unconcernAed about the input; it deals centrally with the output. These two types of constrainAts aDre often in conflict because markedness enforces particular structures on candidates which faithfulness is in conflict with. Examples of markedness constraints include the foIlBlowing: 1. Vowels must noFt be nasal. 2. Syllables must not have codas. 3. Obstruen tsO must not be voiced in coda position. 4. SonorYants must be voiced. 5. Syllables must have onsets. 6. IOTbstruents must be voiced after nasals. RS (Kager, 2001: p. 9) For a cEandidate to be optimal, “some constraint that favors the winner over the loser must dIomVinate every constraint that favors the loser over the winner. The logic of this statement Nfollows from the properties of EVAL. Constraint ranking arguments depend on this logic” U (McCarthy, 2007: p. 7). There are three important elements that validate a ranking argument. First, the ranked constraints have to conflict in their evaluation of competing candidates from an input; second, one of the competing candidates has to be the harmonic candidate and the constraint that favours it must be more highly ranked than that which favours the other candidate(s). Third is that the ranking argument is safe only when there is no third constraint. Constraints are violable, however, violability must be minimal. 60 Optimality Theory neither compromises nor suppresses constraints, instead, what holds is a strict domination of constraints in a hierarchy. In this domination, the more highly ranked constraint is regarded above the other which is not as highly ranked. Therefore, a candidate becomes optimal when it incurs a violation of the least number of highly ranked constraints (McCarthy, 2007). Y Furthermore, a lower ranked constraint can be violated in order to prevent a candidRate from violating a higher-ranked one; violation is often kept to the minimum in orderA to meet the requirement of maximal harmony. The core essence of minimal violationR of constraints is that every instance of violation of constraint should help prevent thIe Bviolation of a higher ranked constraint (Prince and Smolensky, 1993). L 2.4.3 EVAL(uator) N This can be termed the most important component in tAhe theory because it accounts for all “observable regularities of surface forms. EVALD is structured as a (language-specific) hierarchy of universal constraints” (Kager, 20A01: p. 20-21). Constraints hierarchy has a list of the universal constraints and they a reI aBrranged based on their language specificity. The EVAL chooses the optimal candiFdate, that is, the candidate which does not violate the highly ranked constraint and tOhis is chosen as the output. Using the language-specific and -ranked constraints, EVAYL c onsiders the candidates. The candidate which does not violate the highest rankedI cTonstraint is considered to be the optimal candidate. The number of violations is noSt the first or most important consideration, rather, it is the violation or not of the higRhly ranked constraints. The candidate that attracts the smallest number of violatioEns of highly ranked constraints emerges as the optimal candidate. This evaluation oIf Vthe generated candidates is done in parallel form using a table called tableau. NSummarily, GEN generates all the possible candidates, EVAL selects the harmonic one U using CON as the standard for choice (Kager, 2001). 2.4.4 Tableau Tableau is the table used for analyses of the generated candidates; the constraints occupy the first row of the table while the possible outputs, that is, the candidates, occupy the first 61 column. The input occupies the first cell in the first row and column, that is, the uppermost left corner of the table (McCarthy, 2007). There are two types of tableau – the comparative and the violation tableaux; these have their peculiarities and specific functions which distinguish them. In a comparative tableau, each of the cells indicates how often a constraint has been violated and each cell carriYes a W (winner) or an L (loser) to show whether or not a candidate violates a constraRint. In a violation tableau, asterisks (*) are used to indicate violations; asterisks and Aexclamation marks (*!) are used to indicate a Fatal Violation – this is a violation of thRe highest ranked constraint, after which no other evaluation is required as an optimalI cBandidate would have been identified. The ranking of constraints is based on a hierarchyL of strict domination. This situation is called the strictness of domination and it simply means that if a candidate violates only one constraint and this constraint is a highAly rNanked one, then it has performed worse than a candidate which violates many lower-ranked constraints. Hence, an optimal candidate need not obey all the constraints but the Dhighly-ranked one. An icon of a pointing hand () is used to indicate the optimal caBndidAate. Ranking among the constraints is shown by an unbroken line in the tableau and Iby “>>” in prosaic form. Some cells in the tableau are usually shaded to show that theF violations that occur inside these cells are not important. Violation tab leOaux are not as good as comparative tableaux in showing how the constraints and their ranking function in a pIarTticu Ylar language. But they are more useful than comparative tableaux for determining which members of a Sgiven candidate set are possible winners under different R rankings of a given set of constraints. (McCarthy, 2007: p. 8) E BIeVlow are illustrations of what comparative and violation tableaux are as culled from NMcCarthy (2007) U Tableau 2.1 Comparative tableau Tableau 2.2 Violation tableau 62 /Ɂilk-hin/ *Cu DEP /Ɂilk-hin/ *Cu DEP a. → Ɂilikhin 1 a. → Ɂilikhin * b. Ɂilkhin W1 L b. Ɂilkhin *! Y R McCarthy, J. J. 2007. What is optimality theory? Linguistics DepRartmAent Faculty Publication Series. Massachusetts: University of MassIaBchusetts Amherst 27. N L DA IB A F Y O T SI VE R UN I 63 Tableaux 2.1 and 2.2 present two candidates (‘Ɂilikhin’ and ‘Ɂilkhin’) as well as two constraints (*Cu and DEP). Each of these candidates violates one constraint each, however, *Cu dominates DEP because it is more highly ranked and this is shown by its appearance on the leftmost part of the column. Candidate (b)’s violation of the more highly ranked constraint (*Cu) is a fatal one, hence candidate (a)’s violation of DEP does not prevent it from being the harmonic candidate of the tableau. The fatality of candidate (b)’s violaYtion is shown with asterisk and exclamation mark (*!). A closer observation of the tablReau also shows that the last column has shaded portions – this shows that the violations Ain these cells are no longer necessary because a fatal violation which proves the IoBptim Ral candidate has already taken place. L Another tableau presented below to explain some more feature s which are not captured in McCarthy’s (2007) description. DA N A IB F O TY I ER S IV UN 64 Tableau 2.3 Illustrative tableau Input Constraint 1 Constraint 2 Constraint 3 Constraint 4 a. Candidate 1 *! * Y  b. Candidate 2 * AR c. Candidate 3 * * BR Constraints ranking: Constraint 1 >> Constraint 2 >> ConstraiLnt 3I, Constraint 4 Kager, R. 2001. Optimality theory. Cambridge: CNamb ridge University Press. p. 5. DA A F IB O TY SI VE R UN I 65 The styles of line in a tableau are not randomly chosen, instead, each style holds some significance towards the overall analysis in the tableau. First, the single unbroken line between constraint 1 and constraint 2 shows that the first constraint ranks higher than the second. Also, the double and unbroken lines that separate constraints 1, 2 from constraints 3, 4 show that the first two constraints jointly rank higher than the other two constraints. Lastly, the broken line between constraints 3 and 4 shows that neither of them rankRs higYher than the other. Rather, they are on the same level of ranking. These are represented with the use, or absence as the case may be, of ‘>>’ in the prosaic representation ofA the ranking. The shaded portions of the tableau show areas where an activity canInBot im Rpact the overall outcome of the tableau. L 2.4.5 Optimality Theory and phonological processes N McCarthy (1997) reports that Optimality Theory can Abe a verifiable theory to account for phonological processes; he states that process-speDcific constraints are available and these are a result of constraint-ranking – a core prinAciple in the theory. The study counters Davis (1995) which asserts that the effec tsI Bof process-specific constraints are not within Optimality Theory. F Concerning the applicationO of Optimality Theory to children’s language use, Shooshtaryzadeh (2015) ass erts that at the beginning of language development, children simplify their languIaTge Yuse and this prompts the dominance of faithfulness constraints by markedness coSnstraints. RTo produce the correct forms, children should rerank their E initial constraint hierarchy during phonological V development. To rerank a constraint hierarchy, the I markedness constraints is demoted (sent to a lower rank) to N the right side of the faithfulness constraints, i.e. IDENT‑IO. In this way, the child’s language system gradually U approximates to the adults’ language system during language development. (p. 16) Shooshtaryzadeh (2015) goes on to state about children’s error patterns that they are derivations from a hierarchy of conflicting universal constraints. Also, Optimality Theory 66 is endowed with more analytical frameworks to provide more elaborate explanations than the underlying phonological constraints in generative phonology. 2.5 Summary This chapter of the research has presented the conceptual and empirical review of relevYant literature. It also contains a discussion of the theoretical framework that was adopted for the study. In addition, necessary conceptual clarifications were made. The nextR chapter contains the methodology for the research. RA IB N L AD A IB O F Y SI T R IV E UN 67 CHAPTER THREE METHODOLOGY RY 3.0 Preamble A This chapter presents the methodology that guides the different stages oRf the research. It provides details about the design of the research, the samplinIg Btechnique, research instrument, method of data collection, and information about th eL analysis of the data. 3.1 Research design AN The study adopted a descriptive research desigDn. The different types of phonological processes that the children produced were Aperceived and described, theoretically and acoustically. Descriptive statistics was aIlsBo used. 3.2 Sampling technique F The purposive sampling met hoOd was used in the selection of the participants; this is because participants were mainYly selected based on availability, willingness and ability to participate, and theIirT fulfilment of the age criterion (between one and six years old). The age range was aS deliberate choice so as to situate the data collection within the period when the phonolRogical processes, as they relate to children, are still being used. VE3I.3 Sample Size NA total of one hundred schoolchildren were selected from Oyo and Lagos States. These two U states were used because of their proximal, cosmopolitan and multicultural features which provided a relatively broad linguistic milieu for a balanced research. Also, these features meant that dialectal variations in these areas were minimal compared to other states in south-western Nigeria. There were seventy-five (75) children from Lagos State and twenty- five (25) from Oyo State. 68 Also, the utterances of two female children aged one year, three months and four years, three months were sampled for 6 months for a longitudinal observation. This was for the purpose of discovering any language development that occurred over this period of time. The choice of these participants was dictated by their age, availability and an authentication of their medical soundness in terms of speech or cognitive defects. The age range was a purposive choice so as to represent the beginning of utterance production as well as theY end of the use of phonological processes as they relate to children. AR 3.4 Research instrument BR The participants read aloud a purposively designed text which containIed some phonological processes. There were three sections of short and simple stories , aLnd there were two popular rhymes in the text. A total of two hundred and seventyN-five (275) words were in the instrument. Colourful pictures were also insertedD in Athe text to illustrate the stories and rhymes. Words that Nigerian children are foAund to commonly use were deliberately put into the instrument such as grandpa, yo-yo, birthday, cake and so on. This instrument is present in the appendix. IB 3.5 Ethical approval OF The researcher got a letter of introduction from the Head of Department, English and presented it to the adTminYistrators of the schools from which the data were collected. Data collection was ScarrIied out only after the school had consented to the request. Afterwards, the schoolsR presented the researcher with a letter to attest that some children participated in the datEa collection process. The letter of consent and attestation was sought and got from theV school because the data collection occurred during school hours and the administrators wIere made to understand the research and academic objective behind the process. UNIn the case of the longitudinal observation, the parents of the two children consented to allow their children to anonymously participate in the procedure. They wrote and signed a letter of consent to this effect. 69 3.6 Method of data collection The one hundred schoolchildren from whom the data were collected, had acquired a reading skill and were made to read aloud a two hundred and seventy-five word-count text. Their renditions of these were audio-recorded. Y These elicitations and recordings were carried out by the researcher and a research aRssistant to aid a faster process of recording. Prior to the commencement of data collecAtion, a brief but effective training was given to the research assistant on the procRess for the data collection. An average of ten minutes was spent with each child; duriBng this time, each of them was allowed to go through the text before they eventually rLeadI it aloud. An audio-recording application on a mobile phone was useNd to collect the recordings of the participants while they read aloud the prepared texAt and poems. Interfering calls and messages during the recordings were forestalled bDy switching the mobile device to airplane mode. Different rooms were used for the recoArdings to prevent sound overlap. The rooms were quiet in order to limit background InBoise to the barest minimum. For the longitudinal data collectionF, the younger participant who resided in Oyo State was recorded on an average of fourO and a half hours per week. The older participant who resided in Lagos State was recoYrde d for an average of thirty minutes on weekdays. These were audio-recordings oIf Ther utterances while she conversed with her parents and siblings and played with heSr friends. These recordings were done in a way that prevented distraction and put theR participants in a natural state. 3.7 EMethod of data analysis V NTIhree major levels of analysis were adopted in the research – perceptual, theoretical and U acoustic. These were complemented by descriptive statistical analysis. They were combined to provide a result as genuine and authentic as possible. The recordings of the children’s utterances were played back to isolate the different instances of phonological processes. These were grouped into the different types and the frequency of instantiations were noted and converted to percentages. Then, these processes were examined and subjected to analysis using Optimality Theory. The rankings of the different relevant 70 constraints were presented as found in the grammar of the children. Afterwards, representations of the most prominent patterns were subjected to acoustic analysis using version 6.1.40 of Boersma and Weenink’s (2021) PRAAT. The longitudinal data from both participants were transcribed and the converging and unique phonological processes were subjected to theoretical and acoustic analyses. Y 3.8 Summary AR In this chapter, all the details about the design of the research, the samplinRg technique and sample size have been presented. Also, the necessary informatioInB about the research instrument, ethical approval, method of data collection and data analysis has also been included in this chapter. The next chapter contains the data ana lyLsis. N AD A IB OF TY RS I E IV UN 71 CHAPTER FOUR DATA ANALYSIS AND DISCUSSION RY 4.0 Preamble A This section of the research contains the theoretical and acoustic analyses oRf the data which was gathered from the participants. The data were played back and thBe different instances of phonological processes were isolated, grouped according tLo tyIpes, quantified using descriptive statistics and analysed theoretically. N 4.1 Socio-demographic description of the particAipants The tables below show how the participants were Ddistributed based on age, and States. The first table shows the distribution of age, the sAecond shows that of gender, while the third shows their distribution across the two SItBates from where the children were selected. OF Y SI T R IV E UN 72 Table 4.1 Age of the respondents Distribution of age Y Cross-sectional data Longitudinal data R Age Freq (%) Age Freq (R%) A 4 years 19 19.0 1 year, 3 months 1 IB50.0 5 years 31 31.0 4 years, 3 months L1 50.0 N 6 years 50 50.0 A Total 100 100.0 ToAtal D 2 100.0 IB OF Y IT ER S NI V U 73 As presented in Table 4.1, a total of one hundred children from the cross-sectional data and two children from the longitudinal data participated in the study and their ages ranged from one to six years. In the longitudinal data, one child was one year and three months old,Y and the other was four years and three months old. In the cross-sectional data, ninetAeen Rchildren were four years old, thirty-one children were five years old, and fifty wereR six years old. IB L DA N A IB OF TY SI VE R UN I 74 Table 4.2 Distribution of the participants between the states Cross-sectional Longitudinal States Freq (%) Freq (%) (%) Y Lagos State 75 75.0% 1 50.0% 50.0 AR Oyo State 25 25.0% 1 50.0% 5R0.0 B Total 100 100.00 2 TLota Il 100.0 AN BA D I OF TY RS I E NI V U 75 In Table 4.2, the frequency and percentages of the children between the two states is presented. The majority of the participants (73.5%) were from two schools in Ifako and Alimosho Local Government Areas of Lagos State, while twenty-five (24.5%) were from a school in Ibadan, Oyo State. One child each was sampled in Lagos and Oyo States; this constituted 1.0% each. Y 4.2 Phonological processes R Certain structures in the instrument were used to test some processes and theA realisations of these structures have been analysed under their respective titles. TIhBe tes Rted processes are yod coalescence, substitution, monophthongisation, assimLilation, deletion, vowel strengthening and gliding. The other processes found in the d ata are also analysed in this section. N 4.2.1 Coalescence DA Miss Jane, second year, grandma’s yummy anAd last yoyo were the four structures that were purposively inserted into the instrumenIt Bto test the participants’ yod coalescence. In the instrument, Miss Jane occurred 8F tim es, while the rest occurred once only. Thus, while Miss Jane was expected to have a total occurrence of 800, the other structures were expected to have a total oc cuOrrence of 100 each. The frequency and percentages of the realisations of this prToceYss are presented in the table below. SI R EIV UN 76 Table 4.3 Realisation of coalescence in four to six year olds Process Instances Freq ET Variants Frequency % 4 5 6 years years years Coalescence Miss Jane 8 800 /mizʤɛn/ 86 118 305 Y63.6 /misʤɛn/ 57 127 A53R 29.6 /misisʤɛn/ 9 14R 7 3.8 /misʤein/ - IB8 16 3.0 Second year 1 100 /sekɔnjie/ 4L6 13 30 89 /sekɔndjie/ N 1 5 5 11 Grandma’s 1 100 /grDanmAazjɔmi/ 5 30 40 75 yummy A/granmajɔmi/ 18 3 4 25 Last yoyo 1 I10B0 /lastjojo/ 14 39 31 84 /lasjojo/ 8 3 2 13 O F /lastjɔjɔ/ 3 - - 3 A TY 12 1200 /ɛ/ 346 376 406 94 I /ǝ/ - 46 62 6 Key: RS ET: ExEpected Total V UN I 77 Table 4.3 presents the four variants for Miss Jane which were found in the data as produced by the participants. The most frequent variety was /mizʤɛn/ with a total occurrence of 509 (63.6%) out of an expected total of 800. This variety involved a voicing of the coda in the first syllable which can possibly be interpreted as a regressive influence of the voice feature of the voiced palato-alveolar affricate /ʤ/ in the onset of the second syllable. The next variant was /misʤɛn/ with a total of 237 (29.6%). The third variant involveYd a pronunciation of “Miss” as “Mrs”, thus leading to an insertion of /is/ in-betweRen both syllables. Perhaps, this can be interpreted as an oversight on the part of the Aparticipants, especially as only a small 3.8% of them enunciated the structure tBhis Rway. The fourth variant was /misʤein/ and it occurred 24 times in the data, consLtitutIing 3.0% of the data. Thus, the process that can be said to have really occurred hNere w ould be yod dropping since the anticipated coalescence did not occur. This is likAely the case because of the pace at which the participants enunciated the text. AlthouDgh they were given some minutes to go through the text before they eventually read it aloud for audio-recording, they still did so at a relatively slow pace of 4.8 minutes per paArticipant on an average. This meant that the sounds did not flow very well into o nIeB another as is the context for the production of phonological processes. F Second year was the second sOtructure used to test coalescence among the participants and there were two variantYs in the participants’ rendition. These were /sekɔnjie/ with a frequency of 89 anId T/sekɔndjie/ with a frequency of 11. The expected total frequency was 100 and a majoSrity of 89 participants removed voiced alveolar plosive /d/ in the coda of the first syllabRle. The context for an achievement of coalescence involves two different sounds, in this Ecase voiced alveolar plosive /d/ and palatal semivowel /j/, therefore, the removal of thIeV initial phoneme rendered the coalescence unachievable. The other variant (/sekɔndjie/) Nstill retained the two sounds, however, coalescence was not achieved because the words U were read out apart, with a brief pause in-between them, as a result of the slow pace with which the participants read the text aloud. The next structure used to test coalescence was Grandma’s yummy and two variants of this structure were also found in the data, these were /granmazjɔmi/ and /granmajɔmi/ with a frequency of 75 and 25, respectively. There was a voicing of the morphophoneme for 78 possession (-s) in the first variant and its total removal in the second variant. Both variants did not enable coalescence because of their pace and the deletion, respectively. The final structure used to test coalescence was last yoyo with voiceless alveolar plosive /t/ and palatal semivowel /j/ providing the context for this process. The variants that the participants produced were /lastjojo/, /lasjojo/ and /lastjɔjɔ/ with a frequency of 84%, 1Y3% and 3%, respectively. The highest occurring variant did not contain coalescenRce even though the necessary sounds for its occurrence were present. This can be attrAibuted to the care with which majority of the participants read aloud the text despite havRing gone through it once before. The next variant that occurred frequently was /laIsjBojo/ where voiceless alveolar plosive /t/ was removed, thus, eliminating the chanc eL for coalescence. The last variant was like the first, voiceless alveolar plosive /t/ andN palatal semivowel /j/ were still present in the words but, coalescence did not occur. TAhis third variant had 3%. Thus, all four words of the linguistic structures discDussed above did not produce any process of coalescence, because of the careful, emAphasised and discrete enunciation by the participants. However, coalescence occurBred between half-open front spread vowel /e/ and half-close front spread vowel /i/ in elev eIn words in the data. These sounds occurred together as closing diphthongs /ei/ in the wFords in the table, but were rendered as half-close front spread vowel /ɛ/ predominan tOly across these eleven words. Therefore, coalescenTce iYn Nigerian children’s spoken English occurs in diphthongs but not in consonants. TheI articulation proximity of these vowels instigates this process while this is not the case iSn consonants’ articulation despite articulation proximity. 4.2.2 ESubRstitution SIevVen words were purposively used to test the participants’ substitution process and these Nwords were with, the, there, birthday, then, father, and this. With occurred twice in the U instrument and the total expected frequency was 200. The and this, being common words, occurred seven times each in the instrument, and there was an expected total occurrence of 700 tokens in each case. Birthday and father occurred three and two times, respectively, so there was an expected total occurrence of 300 and 200 in that order. Finally, there and then occurred once each, and there 100 tokens of occurrence were expected for each word. The table below presents the information at a glance. 79 RY BR A I N L A AD IBF Y O T SI VE R UN I 80 Table 4.4 Realisation of substitution in four to six year olds Process Instances Freq ET Variants Frequency % 4 years 5 years 6 years Substitution With 2 200 /wit/ 38 56 82 Y88 /wiθ/ - 6 18 R 12 The 7 700 /di/ 133 189 A315 91 /ði/ - 28B R35 9 There 1 100 /de/ 19 L2I7 43 89 /ðe/ - N 4 7 11 Birthday 3 300 /bedɛ/ A48 60 150 86 /bɜ:θdɜD/ - 9 11 6.7 B/beAtdɛ/ 3 6 10 6.3 I /bɜ:θdei/ - - 3 1 Then 1O F100 /den/ 19 30 48 97 Y /ðen/ - 1 2 3 FaIthTer 2 200 /fada/ 37 48 104 94.5 S /fa:dǝ/ 1 3 3 3.7 ER /fa:ðǝ/ - - 4 2.0 IV This 7 700 /dis/ 682 133 203 346 97.4 N /ðis/ 18 - 14 4 2.6 U Key: ET: Expected Total 81 Table 4.4 presents the instances of the words used to test the substitution process. The variants of each of these instances are presented too, along with their frequencies and percentages. For the instance of with, there were two variants: /wit/ and /wiθ/ with frequencies of 176 and 24, and percentages of 88% and 12% in that order. The had an expected total occurrence of 700 and there are two variants of this instance. YThe first is /di/ (637: 91%) and /ði/ which occurs 9% with a frequency of 63. There Ralso had two variants which are /de/ and /ðe/. These variants have a percentage occurrAence of 89% and 11%, and a frequency of 89 and 11, respectively. R In the case of birthday where the expected total occurrence of tokenIsB was 300, there were four variants of the process of substitution and these were /be dLɛ/ (86%), /bɜ:θdɜ/ (6.7%), /betdɛ/ (6.3%) and /bɜ:θdei/ (1%). The frequencies of the Nvariants were 258, 20, 19 and 3, respectively. Unlike the previous instances of words uAsed to test substitution, this word has two syllables and has voiceless dental fricative /θ/ Dand voiced alveolar plosive /d/ occurring side by side. This provided a complex structurAe for the participants to produce and probably explains the higher frequency of variatiIonB, relative to the other instances. The next word was then and thereF we re two variants produced by the participants. These were /den/ (97%) and /ðen/ (3O%). Similarly, the tested word this also had two variants of /dis/ (97.4%) and /ðis/ (Y2.6%). Father had an expected total occurrence of 200 because it occurred twice in thTe instrument. Three variants were found in the data: /fada/ (94.5%), /fa:dǝ/ (3.7%) Sand I/fa:ðǝ/ (2.0%). Voiced anRd voiceless dental fricatives /ð θ/ were mainly used to test the process of substitEution and despite the extremely varied frequencies of the words, the percentages still reIvVeal a particular pattern of substitution. The table shows that voiced dental fricative /ð/ Nwas predominantly substituted with voiced alveolar plosive /d/, while the voiceless dental U fricative /θ/ was mainly replaced with voiceless alveolar plosive /t/. Based on the overall frequencies and percentages in the table, it can be concluded that more Nigerian children substitute voiced dental fricative /ð/ with voiced alveolar plosive /d/, and voiceless dental fricative /θ/ with voiceless alveolar plosive /t/. 82 4.2.3 Monophthongisation A total of seventeen words were used to test the process of monophthongisation among the participants. Gave, Day, Play, Cake, There, Where, Take, Later, and Hair were the nine words used and each occurred once only in the instrument; thus, there was an expected total frequency of 100 tokens per word. Home, Gate, and Away occurred twice in the instrument, therefore, there was an expected total frequency of 200 tokens for each of the three woYrds. Rain occurred 4 times in the instrument, thus, there was an expected total frequencRy of 400 tokens for this instance. Likewise, Rose, Jane, I and A occurred 6, 8, 7 anAd 12 times, respectively in the instrument, so there were expected total frequencBies oRf 600, 800, 700 and 1200, respectively. Table 4.6 below shows the instances, their Ivariants as well as the frequencies and percentages of these variants. N L A AD IB OF Y IT ER S V UN I 83 Table 4.5 Realisation of monophthongisation in four to six year olds Process Instance Freq ET Variants Frequency % 4 years 5 years 6 years Monophthongisation I 7 700 /ai/ 126 170 348 92 /a/ 7 21 28 8 Vine 6 600 /vain/ 104 182 299 97.5 /vin/ 10 4 1 Y 2.5 Time 1 100 /taim/ 19 31 AR50 100 A 12 1200 /ɛ/ 228 336 564 94 /ǝ/ 2 34 36 6 Gave 1 100 /gɛv/ 18 2R8 48 94 /geiv/ - B3 2 5 /gav/ 1 I - - 1 Day 1 100 /dɛ/ 19L 28 47 91 /dei/ N - 3 3 6 Play 1 100 /plɛA/ 19 29 46 94 /plei/ - 2 4 6 Cake 1 100 D/kɛk 19 29 48 96 /keik/ - 2 2 4 There 1 10A0 /de/ 18 26 42 86 IB /ðe/ 1 4 6 11 /ðeǝ/ - 1 2 3 Where F1 100 /we/ 19 30 48 97 /weǝ/ - 1 2 3 Tak e O 1 100 /tɛk/ 18 30 49 97 /teik/ 1 1 1 3 YRain 4 400 /rɛn/ 72 120 196 97 IT /rein/ 2 4 6 3 Home 2 200 /hom/ 38 62 98 99 S /hǝum/ - - 2 1 R Gate 2 200 /gɛt/ 38 58 96 96 /geit/ - 4 4 4 E Rose 6 600 /roz/ 114 186 298 96 IV /rǝuz/ - - 2 4 Jane 8 800 /ʤɛn/ 152 240 384 97 N /ʤein/ - 8 16 3 U Later 1 100 /lɛta/ 19 29 47 95 /leitǝ/ - 1 2 3 /leita/ - 1 1 2 Hair 1 100 /he/ 19 31 48 98 /heǝ/ - - 2 2 Away 2 200 /awɛ/ 38 55 90 41.5 /awei/ - 6 8 7 /ǝwei/ - 1 2 1.5 84 Table 4.5 presents the realisation of the process of monophthongisation in each of the instances of the words used in the instrument. The indefinite article a occurred most frequently in the instrument with a frequency of 1200. Monophthongisation occurred in 1128 instances (94%), reducing the closing diphthong /ei/ to a half-close front spread vowel /ɛ/. This closing diphthong /ei/ was also reduced to only half-open central neutral vowel /ǝ/ in 72 instances of the tokens, constituting 6% of the total occurrence. This substitution Ywas a product of the careful and discrete enunciation style that majority of Athe Rchildren employed. R The first person singular pronoun, I, had a frequency of 7 in the instrIumBent and an expected total frequency of 700 in the entire data. Monophthongisation oLccurred in only 8% of the data with a total frequency of 56 out of 700. The majorityN of the participants retained the closing diphthong /ai/ with a frequency of 644 (92%).A Vine and time also occurred in the instrument; vineD occurred six times, while time occurred once. Thus, there was an expected total of six hAundred and one hundred tokens, respectively for the words. There were two variants fIoBr vine but only one for time. 585 of the tokens for vine were produced as /vain/, wFhile 15 were produced as /vin/, giving a percentage occurrence of 97.5% and 2.5%O, respectively. As for time, all the participants produced it as /taim/, giving 100% for this production. Gave occurred oncIe Tin t Yhe instrument and there was an expected total of 100 tokens in the data. Three varSiants of the word were produced and these are /gɛv/ (94%), /geiv/ (5%) and /gav/ (1%)R. Thus, monophthongisation occurred 95% of the time. The 1% occurrence of /gav/ iEs not significant at all, neither is that of /geiv/ which occurred 5% of the time. DIaVy, play, cake, and take all occurred once each in the instrument and there was an expected Ntotal of 100 tokens each for the words. There were two variants for each of the words: /dɛ/ U and /dei/; /plɛ/ and /plei/; /kɛk/ and /keik/; and /tɛk/ and /teik/. In the case of day, /dɛ/ and /dei/ had a percentage occurrence of 91% and 6%, respectively. In the instance of play, the variants /plɛ/ and /plei/ had an occurrence of 94% and 6% in that order. The variants for cake were /kɛk/ and /keik/ with 96% and 4%, respectively. Take also had two variants, /tɛk/ 85 and /teik/, with respective frequencies of 97% and 3%. All these show a predominance of half-close front spread vowel /ɛ/, a pure vowel that has replaced the closing diphthong /ei/. Where and hair have the same diphthongs and both of them occurred once each in the instrument, so a total of 100 tokens was expected for each of them. Both words contain the centring diphthong /eǝ/, and monophthongisation occurred to convert /weǝ/ to /we/ inY the case of where, and /heǝ/ to /he/ in that of hair. In where, the available variants in Rthe data were /weǝ/ and /we/ with a frequency of 97% and 3%, respectively. A similar vAariation was observed in hair too, where /heǝ/ and /he/ had a frequency of 98% and 2R%, respectively. Thus, monophthongisation was achieved in these diphthongs throIuBgh a removal of the weaker vowel segment, half-open central neutral vowel /ə/. L Home and rose also occurred in the text, home occurred twiNce while rose occurred six times in the text. Thus, a total of 200 and 600 tokens respectivAely were expected in the data. There were two variants in the data – /hom/ and /hǝuDm/ with a percentage of 99% and 1%, respectively. Of the 200 expected tokens, /hAǝum/ was produced only twice in the data. Similarly in rose, there were two vari anItsB – /roz/ and /rǝuz/ with a percentage of 96% and 4%, respectively. Monophthongisation occurred in these cases through merging the vowel sounds in some of the diphthoOngs. For instance, while schwa /ǝ/ is a central vowel, /u/ is a back vowel, the coalition o f these two sounds into one resulted in /o/ whose point of articulation on the toTnguYe is located somewhere between the central and back region. Rain, gate, JanSe, laIter and away all contain the same closing diphthong /ei/. Rain occurred four times Rin the data, thus, there was an expected total frequency of 400 for this token. Gate aEnd away occurred twice each, therefore, there was a total of 200 tokens expected for eIacVh of these words in the data. Jane had a frequency of eight words in the data, meaning Nthat each participant was expected to produce the word eight times, resulting in a total U frequency of 800 tokens in the entire data. Finally, later occurred once only, and there was an expected total of 100 for its frequency in the data. Rain had two variants in the data – /rɛn/ and /rein/ with a percentage of 97% and 3%, respectively. Gate also had two variants with vowel phonemes similar to that of rain; the variants for gate are /gɛt/ and /geit/ and the percentages of their occurrence are 99% and 86 1%, respectively. The frequency of occurrence of the second variant is insignificant with only 8 occurrences out of a total of 200. Jane also had two variants – /ʤɛn/ (97%) and /ʤein/ (3%). Later and away, unlike the others, had three variants each. For the word later, the variants were /lɛta/, /leitǝ/ and /leita/ with occurrence of 95%, 3% and 2%, respectively. With these variants however, the vowel phonemes produced had two variants only, a monophthong in the most frequently occurring variant, and a diphthong in the other Ytwo variants. So as far as monopthongisation is concerned in this token, there were reRally two variants of the vowel phoneme, and they are /ɛ/ (95%) and /ei/ (5%). AwaAy had three variants which are /awɛ/, /awei/ and /ǝwei/ with a percentage of B83%R, 14% and 3%, respectively. For the treatment of monophthongisation, these variantIs can be collapsed into two to give a percentage of 83% for /ɛ/ and 17% for /ei/. TLhus, monophthongisation prevailed in this instance with 83%. N Therefore, monophthongisation in Nigerian childrDen’sA spoken English occurs least in /ai/. However, /ei/, /əu/ and /eə/ are often monopAhthongised and this monophthongisation is executed using other processes: coalescenBce and deletion. I 4.2.4 Assimilation F Four linguistic structures we reO used to test the process of assimilation and these are of blue, like blue, Jane’s home, Yand like this. Of blue, like blue, and Jane’s home appeared once each in the instrumIeTnt, meaning that a total of 100 tokens were expected of them each in the data. Like this, however, occurred five times in the instrument, thus, a total of 500 tokens were expecRted Sin the data. Table 4.7 below shows the instances as well as their variants, frequenEcies and percentages of occurrence. IV UN Table 4.6 Realisation of assimilation in four to six year olds 87 Process Instance Freq ET Variants Frequency % 4 years 5 years 6 years Assimilation Of blue 1 100 /ɔvblu/ 10 18 40 68 /ɔfblu/ 5 13 13 Y31 /ɔblu/ 4 3 1 R 8 Like blue 1 100 /laigblu/ 16 30 4A9 95 /laikblu/ 2 1 BR- 3 /lakblu/ 1 L- I 1 2 Jane’s 1 100 /ʤɛnzhom/ 17 31 50 98 home A2 N - - / ʤɛnhom/ 2 Like this 5 500 /laigdisD/ 52 83 213 69.6 B/laiAkdis/ 28 43 16 17.4 I /laidis/ 15 9 41 13 F Y O T RS I VE UN I 88 Of blue had three variants these are /ɔvblu/, /ɔfblu/ and /ɔblu/. Of the three variants, /ɔvblu/ had the highest occurrence with a percentage of 68%. The variant /ɔfblu/ occurred with a percentage of 31%. /ɔblu/ was the least occurring variant among the participants and this had a percentage of 8%. The first and most frequently occurring variant, /ɔvblu/, had the process of assimilaYtion occurring in the coda of its first syllable, where the voiced bilabial plosive /b/ in tRhe onset of the second syllable regressively influenced the voice of the coda in the fAirst syllable. Thus, this changed voiceless labiodental fricative /f/ to voiced /v/. This pRrocess occurred 68% of the time. In the case of /ɔfblu/, assimilation as in the first vIarBiant did not occur as both syllables were produced slowly and apart from each othe r.L There was an instance of elision of a consonant in the third variant. The elided consonant is the coda of the first syllable – voiced labiodental fricative /v/. So this reNmoved the possibility for an assimilation occurring in this instance. DA The next instance is like blue and it also hadA three variants which are /laigblu/, /laikblu/ and /lakblu/. These had percentages o fI 9B5%, 3% and 2%, respectively. In the instance of /laigblu/, the voiceless velar plosivFe /k/ in the coda of the first syllable (laik) was influenced in terms of voice by voiced bOilabial plosive /b/ which occurred in the onset of the second syllable. Thus, /k/ changed to /g/ in the first syllable through the process of assimilation, specifically, regressiTve aYssimilation. The other variant was /laikblu/ with an insignificant percentage occurreInce of 3%. In this case, the voiceless feature in the coda (voiceless velar plosive /k/) of Sthe first syllable was retained, thus, there was no assimilation in these 3% occurrenceRs. This was likely as a result of a pause between both syllables. The final variant, /lakbluE/, though similar to the second, is missing a vowel so that the diphthong was reduced toI Va monophthong. The percentage occurrence of this variant is quite insignificant (2%), Nmeaning that only 6 out of 100 participants produced it thus. U Jane’s home was the third instance of the linguistic structures used to test assimilation among the participants and it had two variants in the data which are /ʤɛnzhom/ and /ʤɛnhom/. The first variant had an occurrence of 98% while /ʤɛnhom/ had an occurrence of 2%. In the first variant, the morphophoneme for possession (-’s) which occurred between Jane and home was transformed from a voiceless alveolar fricative /s/ to a voiced alveolar 89 fricative /z/. This assimilation is best described as progressive, although, argument can also be made for a regressive one. However, a progressive assimilation is a preferred choice because there is already an existing stance in the literature when it comes to morphophonemes and their contexts, that is, the words they are attached to. The voice of the last phoneme to which they are attached to determines the voice that the morphophoneme adopts. Thus, the voiced feature of alveolar nasal /n/ automaticYally influenced -’s to become voiced. This argument is more established than the existeRnce and influence of the voice feature of glottal fricative /h/ on the morphophonemAe -’s. In the second variant however, the morphophoneme for possession was delBetedR completely; this occurred only 2% of the time. LI The final instance used to test assimilation was like this and it occurred 5 times in the instrument. There were three variants of production and tNhese are /laigdis/, /laikdis/ and /laidis/ with a frequency occurrence of 69.6%, 17.4% Aand 13%, respectively. In the variant /laigdis/, voiceless velar plosive /k/ in the cAoda Dof the first syllable was changed to its voiced counterpart /g/ because of the influence of the voice feature in voiced alveolar plosive /d/ in the onset of the second sIyBllable. This influence was backward, thus, what occurred there was a regressive aFssimilation and this occurred in 348 tokens out of the expected total of 500. The s ecOond variant was /laikdis/ and it had a percentage occurrence of 17.4%. Here, the voicYeless feature of the coda in the first syllable was retained regardless of the presence of IthTe voiced feature of the onset in the second syllable. Finally, the least occurring variaSnt /laidis/ involved a removal of the coda /k/ in the first syllable so that it became an Ropen syllable. Thus, there was no context for assimilation to occur in this variant. HowevEer, this variant was quite insignificant because of its minimal frequency of only 13%. TIhVerefore, it can be inferred that assimilation in Nigerian children’s spoken English occurs Nboth regressively and progressively. Also, voice is a predominant feature that is used in this U process of assimilation. 4.2.5 Deletion 90 There were nine linguistic structures used to test the process of deletion and these are next, set, grandpa, grandma, and, friends, combed, want, and textbooks. It is necessary to mention that the process of deletion in some of these words were effective because of the contexts in which they occurred, for instance, and, set and want. That is, the words they occurred with made deletion occur in them. Set and and occurred twice each in the instrument, thus, 200 tokens were expected for each of these words in the entire data. YThe other items occurred once each, meaning that 100 tokens were expected of them eacRh. Table 4.8 below shows the instances, their variants, frequencies and theR perAcentages of occurrence. IB L DA N A F IB O TY SI VE R UN I 91 Table 4.7 Realisation of deletion in four to six year olds Process Instance F ET Variants Frequency % 4 years 5 years 6 years Deletion Next 1 100 /neks/ 11 26 47 84 /nes/ 5 3 1 9 Y /nest/ 3 1 1 5 R /nekst/ - 1 1 A2 Set 2 200 /set/ 31 54 93 R 89 /sed/ 3 5 IB7 7.5 /se/ 4 3 L - 3.5 Grandma 1 100 /granma/ 19 AN31 50 100 Grandpa 1 100 /granpa/ 1D9 31 50 100 And 5 500 /an/ A42 93 190 65 /a nId/B 53 62 60 35 Friends 1 100 F/frendz/ 4 15 39 58 O /frens/ 10 10 9 29 TY /frend/ 5 6 2 13 CoSmbeId 1 100 /komd/ 17 31 50 98 R /komb/ 2 - - 2 VE Want 1 100 /wɔn/ 11 29 49 89 I /wɔnt/ 8 2 1 11 UN Textbooks 1 100 /tezbuk/ 9 19 40 68 /tezbuks/ 7 4 4 15 /tekstbuks/ - 5 6 11 /teksbuk/ 3 3 - 6 92 Next had four variants and these are /neks/, /nes/, /nest/ and /nekst/. It is likely that there were up to four variants because of the number of consonants in the coda of this syllable. The percentages of occurrence of these variants were 84%, 9%, 5% and 2%, respectively. Very few of them (2%) were able to retain all three consonants in the coda of the word, while 9% of them deleted three of these consonants and 5% of them deleted two of these consonants. The most significant variant of all is /neks/ which has the highest percentYage occurrence (84%). Here, the final consonant which is voiceless alveolar plosive R(/t/) was deleted. In the instrument, the word that occurred after next was to; this wordA also begins with a voiceless alveolar plosive. Thus, juncture prosody must be resBponsRible for the high rate of /t/ deletion that occurred in this word. LI The second linguistic item was set and it occurred twice in the instrument, meaning that a total of 200 tokens of this structure were expected in theN whole data. Like the linguistic structure above, that is next, this word also occurred clAose to a consonant. In both cases, set occurred before the voiced dental fricative /ð/. InD this linguistic structure however, there was only 3.5% of deletion and this involvBed thAe voiceless alveolar plosive (/t/) in the coda, while 89% of the tokens retained th isI consonant. Also, 7.5% of these tokens replaced voiceless alveolar plosive /t/ with Fits voiced counterpart. The probable explanation for this lies in the context in which thOis word set occurs. Following it, in both cases, was the word the /ði/ which was enuncYiate d as /di/ 91% of the time. Thus, the occurrence of the voiceless alveolar plosive /d/I bTefore the voiceless alveolar plosive /t/ can be seen as responsible for the substitutionS or assimilation of /t/ to /d/ in these 7.5% cases. The next twRo words, grandpa and grandma, are apparently words that the participants were very faEmiliar with as all produced it the same way – deleting the silent letter ‘d’. Thus, there wIaVs a 100% deletion of /d/ in both cases. UNThe next word, and, occurred five times in the instrument. There were two variants of production in the data and only one of the variants involved a deletion of the final consonant – voiced alveolar plosive /d/, while the other retained it. Thus, /an/ had an occurrence of 65% while /and/ had an occurrence of 35%. Deletion mostly occurred in the instances where a consonant phoneme occurred after and, while /d/ was retained mostly where vowel phonemes occurred after the word. This observation occurred for most, not all of the 93 scenarios with the participants. The consonant phonemes that occurred after and in the data were voiced palato-alveolar affricate /ʤ/, voiced velar plosive /g/ and voiced bilabial plosive /b/. The vowel phonemes that occurred after this token were closed diphthong /ai/ and close front spread vowel /i/. Friends was also used to test the process of deletion; this monosyllabic word has Ytwo consonant phonemes in its onset and three consonant phonemes in its coda pRosition. Deletion only occurred in the coda, producing three different variants: /frendzA/, /frens/ and /frend/ where at least two consonant phonemes were present in each variaRnt, also, alveolar nasal was the constant consonant phoneme that was present in allI tBhree variants. These variants had the percentage occurrence of 58%, 29% and 13%, rLespectively. Like grandpa and grandma, combed had a clearly prefeNrred variant of /komd/, with a percentage of 98%, while the other variant /komb/ hadA a percentage of 2%. This variant in the minority occurred as a result of the participDants’ removal of the regular past tense morpheme, ‘-ed’, which was phonologicallyA represented by voiced alveolar plosive /d/. Want occurred once in the data and hadI tBwo variants in the participants’ production which were /wɔn/ and /wɔnt/ with 89% and 1 1% occurrence, in that order. Thus, there was a high percentage of deletion in this OwordF by the participants. The final word used to Ytest deletion process was textbooks and there are two syllables in this word. The firsIt Tsyllable has one consonant phoneme in its onset and three consonant phonemes in itSs coda, while the second syllable has one consonant phoneme in its onset and two coRnsonant phonemes in its coda. Deletion affected the codas of the two syllables while Etheir onsets were intact. There were four variants of this structure and these are /tIezVbuk/ (68%), /tezbuks/ (15%), /tekstbuks/ (11%) and /teksbuk/ (6%). The most Nprominent variant is /tezbuk/ with 68% and it involved a reduction of the consonant U phonemes in the codas to only one each. Also, the consonant phoneme in the coda of the first syllable changed from voiceless alveolar fricative /s/ to its voiced counterpart /z/ and the latter was retained instead as the sole consonant phoneme in the first coda of the word. It is important to note that this instance shows the deletion of another morphophoneme, this time that of plurality. Generally, the actual total of participants who deleted morphophonemes in this token was 74%. In summary, Nigerian children used deletion 94 process to permit juncture prosody, reduce consonant clusters and get rid of some morphophonemes in their spoken English. 4.2.6 Vowel strengthening The three central vowels /ɜ: ǝ ʌ/ were used to test the process of vowel strengthening, and the following words were employed to this effect: sun, water, bud, cut, up, mum, us, Your, month, little, away and birthday. All these linguistic items appeared once AeacRh in the instrument, except our which appeared five times, away, which appReared twice and birthday which appeared three times. Table 4.9 below shows the linIgBuistic items and their variants, as well as their frequencies and percentages of occurrenLce. N AD A IB F O TY SI ER V UN I Table 4.8 Realisation of vowel strengthening in four to six year olds 95 Process Instance Freq ET Variants Frequency % 4 years 5 years 6 years Vowel Sun 1 100 /sɔ:n/ 19 30 48 97 strengthening /sʌn/ - 1 2 3 Water 1 100 /wɔta/ 19 31 48 Y98 /wɔtǝ/ - - 2 R2 Bud 1 100 /bɔd/ 14 29 48 91 /bud/ 5 2 - A 7 /bʌd/ - - R2 2 Cut 1 100 /kɔ:t/ 19 30I B 48 97 /kʌt/ - L1 2 3 Up 1 100 /ɔp/ 19 N 31 48 98 /ʌp/ - - 2 2 Mum 1 100 /mɔm/ DA19 29 48 96 /mʌm/ - 2 2 4 Us 1 100 /ɔsA/ 19 29 48 96 B/ʌs/ - 2 2 4 Our 5 F50 0 I /awa/ 95 135 230 80 /auǝ/ - 20 20 20 Month 1O 100 /mɔ:nt/ 19 28 42 89 Y /mʌnt/ - 3 6 9 /mʌnθ/ - - 2 2 LIittTle 1 100 /litu/ 17 27 48 91 S /litul/ 3 3 - 6 R /litǝl/ - 1 2 3 E Away 2 200 /awei/ 2 1 4 3.5 IV /awɛ/ 36 61 92 94.5 N /ǝwei/ - - 4 2 Birthday 3 300 /bedɛ/ 51 75 132 86 U /bɜ:θdɜ/ - 7 13 6.7 /betdɛ/ 6 10 3 6.3 /bɜ:θdei/ - 1 2 1 Sun, cut, up, mum, and us all occurred once each in the data and all had two variants each. There is more uniformity to these words than just the number of their variants. They all 96 contain open central neutral vowel /ʌ/ which was replaced by an open back rounded vowel /ɔ/. Although they all had varying percentages, the clearly predominant variants contain /ɔ/ all through. For instance, the variants of sun are /sɔn/ (97%) and /sʌn/ (3%); the variants of mum are /mɔm/ (96%) and /mʌm/ (4%). Bud also contains the open central neutral vowel /ʌ/ and appeared once in the instrumYent, but three variants of this item were found in the data. The variants are /bɔd/ (91%R), /bud/ (7%) and /bʌd/ (2%). It appeared that a few of the participants were unfamilAiar with this word bud, and this was detected because of their longer pause before the pRroduction of this word. Thus, it is proposed that the unfamiliarity with the word mayI hBave prompted 7% of participants to replace the peak in the word with a close back ro uLnded vowel /u/, instead of the open central neutral vowel /ʌ/. Considering this explaNnation, it is proposed that /ʌ/ is generally strengthened till it becomes /ɔ/ because 91%A of the participants did this. Our /auǝ/ is a triphthong which ends in schwa. ODur occurred five times in the instrument, thus, there was a total of 500 expected toBkensA in the data. There were two variants of this item and these are /awa/ (80%) and /aIuǝ/ (20%); thus, the third vowel phoneme in the triphthong, open central neutral vFowe l /ə/, was strengthened till it became an open front vowel (/a/). O The open central neutralY vowel /ʌ/ in month was strengthened till it became an open back rounded vowel /ɔ/ aInTd this occurred 89% of the time in the data. Although there were three variants producSed by the participants, the other two variants differed only in the aspect of the coda, thReir peaks were retained and un-strengthened. LitVtle aElso had three variants, however, as far as vowel strengthening is concerned, there wIere really only two variants which involved the close back rounded vowel /u/ and the Nclose central neutral vowel 11 /ǝ/. Thus, the variants and their percentages are /litu(l)/ 97% U and /litǝl/ 3%. Away appeared twice in the instrument and had three variants in the data which are /awei/, /awɛ/ and /ǝwei/. Their percentages of occurrence were 94.5%, 3.5% and 2%, respectively. So technically, and where it concerned the current process, vowel strengthening occurred 98% of the time. 97 Birthday occurred three times in the instrument, thus, there was an expected total of 300 tokens in the data. There were four variants and these are /bedɛ/ (86%), /bɜ:θdɜ/ 6.7%), /betdɛ/ (6.3%) and /bɜ:θdei/ (1%). In all, 92% contained strengthened vowels which changed from half-close central neutral vowel /ɜ:/ to half-open central neutral vowel /e/. Therefore, it can be concluded that Nigerian children strengthened their central vowYels. Specifically, half-open central neutral vowel /ǝ/ was strengthened into open fronRt spread vowel /a/; half-close central neutral vowel /ɜ:/ was strengthened into half-openA front spread vowel /e/, and open central neutral vowel /ʌ/ was strengthened to open bacRk rounded vowel /ɔ/. IB 4.2.7 Gliding L Gliding occurred mainly in the word our which occurred fiNve times in the instrument, thus, a total of 500 tokens existed for this word in the dAata. Two variants of this linguistic structure were produced and they are /awa/ and D/auǝ/ with frequencies of 80% and 20% respectively. In other words, gliding occurredA 80% of the time where close back rounded vowel /u/ glided into the bilabial semi voIwBel /w/. 4.3 Constraints ranking for pFrocesses O This section of the chapYter c ontains the Optimality theoretical analysis of the instances of the processes that haTve been discussed in the previous section. The constraints rankings responsible forS eacIh emergence are presented in the different tableaux. 4.3.1 CoRnstraints ranking for coalescence FivVe sEtructures were used to test yod-coalescence among the participants and these stIructures were Second year, Grandma’s yummy, last yoyo, a and Miss Jane. All the Nstructures occurred once only except a and Miss Jane which occurred 8 and 12 times, U respectively. Thus, a total of 100 tokens were expected of all the structures except a and Miss Jane, for which a total of 1,200 and 800 tokens, respectively, were expected. Coalescence was predominantly dropped by the participants, and Second year is theoretically analysed in the next paragraph. 4.3.1.1 Realisation of second year 98 In the production of second year, there were two variants discovered in the data and these are /sekɔnjie/ with a percentage of 89% and /sekɔndjie/ with a percentage of 11%. The Received Pronunciation (RP) variant of the utterance and its variants among the participants have been adopted as the candidates for analysis in Tableau 4.1 below. Y R A IB R L N AD A IB OF TY SI VE R UN I 99 Tableau 4.1 Emergence of /sekɔnjie/ “second year” Second year /sekɔnd jie/ UNIF *COMPLEX MAX a. /sekɔnd jie/ * RY b. /sekɔn jie/ * RA c. /sekǝnʤie/ *! IB L AN AD IBF Y O IT RS IV E UN 100 Constraints ranking: UNIF » *COMPLEX » MAX UNIFORMITY: No coalescence (It penalises segmental coalescence. No element of S2 has multiple correspondents in S1) (Wheeler, 2005: p. 62) *COMPLEX: A syllable has no more than one consonant in a margin (Wheeler, 2007:3) MAX: No deletion (Every phoneme in the base should have a correspondent in the output) (Wheeler, 2007:3) RY There are three candidates and three constraints in the tableau. Two of the caAndidates are variants from the data while the third is an RP variant which contains a yRod-coalescence. The first constraint (UNIFORMITY) is the highest ranked in the tableIauB and it prohibits any form of segmental coalescence. Candidate (c) violates this conLstraint by merging voiced alveolar plosive /d/ with palatal semivowel /j/ into a coNales ced voiced palato-alveolar affricate /ʤ/. Because constraint 1 is the highest rankeAd, this violation is a fatal one and the option for a harmonic candidate shifts to either ofD candidates (a) or (b). Constraint 2 prohibits the presence of more thAan one consonant in the margin of a syllable; thus, the presence of alveolar nasal /n / IaBnd voiced alveolar plosive /d/ in the coda of the second syllable in candidate (a) violates this constraint. Therefore, candidate (b) emerges as the optimal candidate becaOuse Fof these violations by candidates (a) and (c). Thus, the cells under constraint 3 are shaded grey because regardless of the violations that occur under this constraintT, thYe emergence of candidate (b) cannot be affected. In other words, candidate (b)’s vioIlation of the third constraint (MAX) does not deter it from maintaining its position as tShe harmonic candidate. 4.3.1.2E CoRalescence of /e/ and /i/ CIoValescence occurred in two adjacent sounds half-open front spread vowel /e/ and half-Nclose front spread vowel /i/. These phonemes occurred together as diphthongs in one tested U word (a), as well as some other nine words in the instrument: gave, day, play, cake, take, rain, gate, later and away. In the production of these structures, there were significant percentages of coalescence in the diphthongs. A and gave had 94% of coalescence each. Day, play and cake had 91%, 94% and 96% of coalescence, respectively. Take and rain both had a 97% occurrence of coalescence in their diphthongs. Gate, later and away had 101 96%, 97% and 95% of coalescence, in that order. The indefinite article “A” is analysed in the tableau below. RY BR A L I AN BA D F I O TY RS I VE UN I 102 Tableau 4.2 Emergence of /ɛ/ “A” A /ei/ *COMPLEXVOW *SCHWA UNIF Y a. /ei/ *! AR b. /ɛ/ * BRI c. /ǝ/ * L AN AD IBF Y O T SI VE R UN I 103 Constraints ranking: *C VOWOMPLEX » *SCHWA » *COMPLEX *C VOWOMPLEX : No string of vowels within a syllable (Green, 2010: p. 101) *SCHWA: the realisation of a central vowel is banned (Wheeler, 2007: p. 8) UNIFORMITY: No coalescence (No element of S2 has multiple correspondents in S1) (Wheeler, 2005: p. 62) Y The tableau contains three candidates and constraints each. The highest ranked coRnstraint is *C VOW OMPLEX and it prohibits the occurrence of complex vowels, that isA, diphthongs and triphthongs. In this case, there is a diphthong in the first candidate, thRus, candidate (a) has violated constraint 1, and this violation is a fatal one. Also, conIsBtraint 2 prohibits the realisation of a central vowel and the presence of half-open ceLntral neutral vowel /ǝ/ as candidate (c) violates this constraint. These violations leaNd to the emergence of candidate (b) as the optimal candidate. This emergence is not prAevented by candidate (b)’s violation of the third constraint which prohibits any formD of segmental coalescence. However, coalescence is exactly what has occurred in Acandidate (b) where half-open front spread vowel /e/ and half-close front spread voIwBel /i/ have merged into one segment /ɛ/. 4.3.1.3 Representative acoustic aFnal ysis of coalescence in /ɛ/ To illustrate coalescence in O/ɛ/, three PRAAT spectrogram images have been used to present the formant freqYuen cies of the two vowel sounds (/e/ and /i/) which merged into a single one (/ɛ/). ThIesTe images are presented and discussed below. S ER IV UN 104 Y AR R LI B N A AD IB F Fig. 4.1 Acoustic illustratio nO of /e/ in /geit/ by a six-year-old TY RS I IV E UN 105 In the acoustic illustration of half-open front spread vowel /e/ above, the four different formant frequencies of the sound are presented, however, only the average of the first two formants are relevant to this analysis and these are 602.26Hz and 1675.35Hz for the first and second formants, respectively. The first formant of this phoneme is conspicuously higher than that of the next sound (half-close front spread vowel /i/) and this presents evidence for a more open articulation of the current vowel phoneme /e/. The lesser vYalue for the second formant frequency in /e/ presents proof that place of articulation iRs not as close to the front of the mouth as it is in /i/. One of the participants whose pAroduction of this vowel segment was representative enough of the majority of the BprodRuctions found in the data was used for this acoustic analysis. The next figure preseLnts Ian acoustic illustration of her production of /i/. DA N IB A OF TY SI VE R UN I 106 Y AR R BLI N AD A IB Fig 4.2 Acoustic illustration of /iF/ in /geit/ by a six year old O IT Y ER S IV UN 107 The figure presents the formant frequencies for half-close front spread vowel /i/, where the average of the first formant frequency is 392.58Hz and the average of the second formant frequency is 2839.09Hz. These averages were automatically generated for approximate accuracy. The low value of the first formant frequency shows that articulation is more closed than open, while the very high second formant frequency shows that the articulation of the vowel occurred at the frontal region of the tongue. RY The average of the first and second formant frequencies of both segments are pAresented and plotted on an acoustic vowel chart below to show their places on the chaRrt as well as the occurrence of the coalescence between both sounds. To situate theI cBoalesced segment on the acoustic chart, its first and second formant frequencies wer e Lcollected and used. This is presented in figure 4.3 below. The first and second formant frequencies for /ɛ/ are 472.63Hz and 2359.92Hz, respectively. Along with the formant valueNs for /i/ and /e/, these values are also plotted on the chart below. AD A F IB O Y IT RS VE UN I 108 F2 [Hz] Y200 250 AR 300 350 R 400 i B I 450 ɛ L 500 550 e N 600 A 650 700 AD 750 800 IB F1 [Hz] Fig. 4.3 Acoustic vowel chart forF coalescence in /ɛ/ by a four-year-old O Y IT ER S V UN I 109 2800 2700 2600 2500 2400 2300 2200 2100 2000 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 900 800 The segments have been plotted unto the acoustic vowel chart using their first and second formant frequencies and from the positions of the three segments. It is conspicuous that half-close front spread vowel /i/ and half-open front spread vowel /e/ have undergone a movement towards each other to form a new sound which is central to both of them. This sound is /ɛ/ and its first and second formant frequencies are located somewhere between those of /i/ and /e/. The arrows from both sounds have been used to portray a more vYivid representation of the coalescence. It is also necessary to state that /i/ as producedR here is much more frontal than found in the English or cardinal vowel chart. RA 4.3.2 Constraints ranking for substitution IB For the process of substitution, the research instrument was uLsed to test voiceless and voiced dental fricatives (/θ/ and /ð/) in different parts of aN syllable and word. In birthday, voiceless dental fricative (/θ/) is located in a coda witAhin a disyllabic word, while in with, it is located in the coda of a monosyllabic wordD. Also, voiced dental fricative (/ð/) was tested in the onset of a monosyllable, the, andA in an onset within a disyllabic word, father. This is the motivation for the theoretic aIl Banalysis of these four instances in the subsequent paragraphs and tableaux. F Also, one instance of the pr ocOess of substitution has been presented from the longitudinal data and this is thing. This instance contains voiceless dental fricative /θ/ in the onset of its syllable and has beIenT the Yoretically analysed in this section as well. 4.3.2.1 RealisaStion of birthday The thEeoreRtical explanation for the emergence of birthday /bedɛ/ in the data is produced bIelVow. This variant had a percentage occurrence of 86%; this makes it the predominant Nvariant in the data. The other variants were /bɜ:θdɜ/, /betdɛ/ and /bɜ:θdei/ with percentage U occurrences of 6.7%, 6.3% and 1%, respectively. Thus, the emergence of /bedɛ/ is analysed in the tableau below. 110 Tableau 4.3 Emergence of /bedɛ/ “birthday” Birthday /bɜ:deɪ/ *SCHWA NODIPHTHONG *C]σ *FRICATIVE MAX a. bɜ:dei *! *! *! * RY b. bɜ:tdeɪ *! *! *! RA c. betdei *! *! L I B d. betdɛ *N! * A  e. bedɛ D * A IB F O Y IT ER S V UN I 111 Constraints ranking: *SCHWA » NODIPHTHONG » *C]σ » *FRICATIVE » MAX *SCHWA: the realisation of a central vowel is banned (Wheeler, 2007: p. 8) NODIPHTHONG: diphthongs are prohibited (Green, 2010: p. 101) *C]σ: Syllables must not have codas (Borowsky, 1989: p. 146) *FRICATIVE – Segments may not be [+cont, –son] (Pater and Barlow, 2003: p. 492) MAX – No deletion (Every phoneme in the base should have a correspondent in theR outYput) (Wheeler, 2007: p. 3) A There are five constraints in Tableau 4.3, and these constraints affecRt the candidates unequally because the constraints rank differently based on the IleBvel of the prevalent hierarchy in the participants’ phonology. The unbroken lineLs between the constraint columns show the hierarchy that exists among these consNtrain ts. GEN has generated five candidates for evaluation, and four of these are varAiants of what was produced by the participants. The first two candidates (a) and (bD) violate *SCHWA, the first constraint, because of the presence of half-close centrAal neutral vowel (/ɜ:/) in them. These two candidates also violate constraints 2 and B3 which are NODIPHTHONG and *C]σ because of the presence of diphthongs and codas wIhich the constraints prohibit, respectively. These violations are fatal and so this ruleFs out the possibility of emergence of either of these two candidates as an optimal ca ndOidate. Candidate 1 also violates constraint 4 because of the presence of a fricative inY it, particularly voiceless dental fricative /θ/ which the constraint prohibits. IT Furthermore, cSandidate (c) violates constraints 2 and 3 because of its retention of a diphthong Rand a coda, respectively. Candidate (d) violates constraints 3 and 5 as a result of theV preEsence of a closed syllable and the absence of a segment that had been there in the inIput phase. Constraint 5 prohibits the deletion of any segment, thus, candidates (d) and (e) Nhave violated this constraint. Candidates (a) to (d) have fatally violated at least one U constraint, while candidate (e) has incurred the violation of a lowly ranked constraint. Thus, candidate (e) emerges as the harmonic candidate in the tableau. In this candidate, there is no diphthong, central vowel or coda as dictated by the higher ranked constraints in the participants’ phonological grammar. 112 This pronunciation of birthday as /bedɛ/ is not unique to these participants; rather, it is a typical pronunciation found among some Nigerian speakers of English. This shows that the participants already acquired their ambient language to a great extent. Another instance of substitution is analysed in Tableau 4.4 below. 4.3.2.2 Realisation of with Y The Optimality theoretical analysis for the realisation of with as /wit/ is carrieRd out in Tableau 4.4 below. Two variants emerged from this structure and these are /wAit/ and /wiθ/ with percentage occurrences of 88% and 12%, respectively. Thus, the eRmergence of the predominant variant is analysed in the tableau below. These twLo Iva Briants constitute the generated candidates in the input of the tableau. N A D A IB O F IT Y ER S V UN I 113 Tableau 4.4 Emergence of /wit/ “with” With /wɪθ/ *FRICATIVE IDENT(PLACE) IDENT(MANNER) Y a. /wiθ/ *! AR b. /wit/ * * R LI B N AD A IBF Y O SI T VE R I UN 114 Constraints ranking: *FRICATIVE >> IDENT(PLACE), IDENT(MANNER) *FRICATIVE: Segments may not be [+cont, –son] (Pater and Barlow, 2003: p. 492) IDENT(PLACE): Assign one violation mark for every place feature in the input that is different from its corresponding place feature in the output (Lamont, 2015: p. 54) IDENT(MANNER): Assign one violation mark for every manner feature in the input that is different from its corresponding manner feature in the output (Lamont, 2015: p. 54R) Y The tableau has two candidates and three constraints. The first constraint (*AFRICATIVE) prohibits the occurrence of any fricative in a structure; thus, the presenceR of voiced dental fricative (/θ/) in the coda of candidate 1 violates this very high ranBking constraint. This single violation leads to the emergence of the second candidateL as Ithe optimal candidate. This explains the grey shading in the rest of the tableau. It sh ows that regardless of what violations or obedience to other constraints that happeAn in Nthis grey part of the tableau, the emergence of the optimal candidate can no lonDger be affected. The ranking of these constraints also shows that constraints 2 andA 3 occupy the same level of hierarchy while constraint 1 is ranked higher than themB both. This is represented by the double lines between constraints 1 and 2, and the do tIted lines between constraints 2 and 3. The presence of >> between the first and secondF constraints as well as the lack of it between the second and third constraints in the proOsaic representation of the constraints is an indication of this hierarchy. Thus, despite th e fact that the optimal candidate violates the other two constraints, it still IreTtain Ys its optimal status in the tableau. Candidate 2 is able to prevent violation of thSe high ranking constraint by changing its fricative to a plosive which is permitted iRn the ranking of these constraints. GeneraElly, the fricative in this input is one that is often substituted even in these pIarVticipants’ ambient language. Therefore, the reranking of constraints in favour of a Nplosive instead of this fricative can be said to be motivated, not just by difficulty, but also U by imitation. 4.3.2.3 Realisation of the In the instrument, four of the linguistic structures used to test substitution had similar structures as well as the same consonant phoneme in their onsets; therefore, only one of them has been analysed theoretically. These words are the, there, then, and this. Also, the 115 same voiced alveolar plosive /d/ was used to substitute the voiced dental fricative /ð/ in these words. Thus, the Optimality theoretical analysis for the realisation of the as /di/ is carried out in Tableau 4.5 below. There are two variants for this structure and these are /di/ and /ði/ and the percentage of their occurrences are 91% and 9%, respectively, thus, the emergence of the predominant variant is analysed in the tableau below. The two variants are part of the generated candidates in the input of the tableau. RY A BR L I AN BA D F I O TY RS I VE UN I 116 Tableau 4.5 Emergence of /di/ “the” The /ðǝ/ *FRICATIVE *SCHWA IDENT(PLACE) IDENT(MANNER) a. /ðǝ/ *! *! RY b. /dǝ/ *! * A* R c. /ði/ *! IB d. /di/ * LN * A BA D F I O TY SI VE R I UN 117 Constraints ranking: *FRICATIVE, *SCHWA >> IDENT(PLACE), IDENT(MANNER) *SCHWA: The realisation of a central vowel is banned (Wheeler, 2007: p. 8) *FRICATIVE – Segments may not be [+cont, –son] (Pater and Barlow, 2003: p. 492) IDENT(PLACE): Assign one violation mark for every place feature in the input that is different from its corresponding place feature in the output (Lamont, 2015: p. 54) IDENT(MANNER): Assign one violation mark for every manner feature in the input thaYt is different from its corresponding manner feature in the output (Lamont, 2015: Ap. 54R) The tableau contains a substitution of voiced dental fricative (/ð/) withR voiced alveolar plosive (/d/). GEN generates four candidates and these are evaluateIdB using four different constraints. Constraints 1 and 2 are more highly ranked than cLonstraints 3 and 4 in the tableau. Also, while constraints 1 and 2 are on the same Nlevel of hierarchy, constraints 3 and 4 are also on the same level of hierarchy. These ranAkings are shown by the broken lines between constraints 1 and 2, and constraints 3 anDd 4. Thus, the violation of constraints 1 and 2 by candidate (a) to (c) prove to be fataAl because both constraints are ranked on the same level in the tableau. These candidatBes contain fricative, schwa or both, and these are the conditions that both constraints pro hIibit. These violations leave candidate (d) to emerge as the harmonic candidate despite Fits violation of constraints 3 and 4 which are faithfulness constraints. These constrain tsO prohibit any kind of change of feature or phoneme between the input and output. Generally also, theIseT Y phonemes (/ð/ and /ǝ/) have proven to be quite problematic for both young and aduSlt speakers of English in Nigeria; thus, it is no surprise that the constraints which EprohRibit these phonemes are ranked highly by the participants. 4I.3V.2.4 Realisation of father NThe last word used to test substitution is father and although it also contains the voiced U dental fricative /ð/, this phoneme occurs word-medially. It is analysed in a tableau to establish the constraints ranking responsible for its emergence as /fada/. There were three variants of father in the production of the participants and these are /fada/, /fa:dǝ/ and /fa:ðǝ/; their percentages of occurrence are 94.5%, 3.7% and 2.0, respectively. These three variants are presented as candidates in the tableau below. 118 Tableau 4.6 Emergence of /fada/ “father” Father /fa:ðǝ/ *SCHWA IDENT(PLACE) IDENT(MANNER) *FRICATIVYE a. /fa:dǝ/ *! * * A*R b. /fada/ * * BR * I c. /fa:ðǝ/ *! * L * N AD A IB OF Y SI T R IV E UN 119 Constraints ranking: *SCHWA >> IDENT(PLACE) >> IDENT(MANNER) >> *FRICATIVE, *SCHWA: The realisation of a central vowel is banned (Wheeler, 2007: p. 8) IDENT(PLACE): Assign one violation mark for every place feature in the input that is different from its corresponding place feature in the output (Lamont, 2015: p. 54) IDENT(MANNER): Assign one violation mark for every manner feature in the input that is different from its corresponding manner feature in the output (Lamont, 2015: p. 54) Y *FRICATIVE – Segments may not be [+cont, –son] (Pater and Barlow, 2003: pA. 492R) There are two candidates and four constraints in Tableau 4.6. The first canRdidate has fatally violated constraint 1 which prohibits central vowels. The other three cIoBnstraints are violated by both candidates but these constraints are not as highly raLnked as *SCHWA so the violations are not fatal. N Constraints 2 and 3 are faithfulness constraints that reAquire faithfulness of output to input in terms of place and manner of articulation. TAhe rDeplacement of voiced dental fricative /ð/ with voiced alveolar plosive /d/ in both candidates means that they have both violated these constraints. However, these constrain tsI Bare not as highly ranked as constraint 1, so the candidates’ violation of them is noFt fatal. The fourth constraint, *FRICATIVE, prohibits the presence of a fricative in a structure and since both candidates have voiceless labiodental fricative /f/ as the consona nOY t in their onsets, they have both violated this constraint. However, like constrTaints 2 and 3, this constraint is not highly ranked, so its violation by both candidates in Ithe tableau does not affect the emergence of candidate 2 as the optimal candidate. RS 4.3.2.5E Realisation of thing TIhVis word has been introduced from the longitudinal data and its production presents an Nillustration of the substitution of voiceless dental fricative /θ/ with voiceless alveolar U plosive /t/ in the onset of the syllable. This process is analysed theoretically in the tableau below. 120 Tableau 4.7 Emergence of /tin/ “thing” Thing /θin/ *F-ONS *FRICATIVE IDENT(PLACE) IDENT(MANNEYR) a. /θin/ *! * A R b. /tin/ * BR * I L N AD A F IB Y O IT ER S IV UN 121 Constraints ranking: *F-ONS » *FRICATIVE » IDENT(PLACE), IDENT(MANNER) *F-ONS: Fricatives are prohibited from occupying an onset position in a syllable (Pater and Barlow, 2003: p. 492) *FRICATIVE: Segments may not be [+cont, –son] (Pater and Barlow, 2003: p. 492) IDENT(PLACE): Assign one violation mark for every place feature in the input that is different from its corresponding place feature in the output (Lamont, 2015: p. 54) Y IDENT(MANNER): Assign one violation mark for every manner feature in the AinpuRt that is different from its corresponding manner feature in the output (Lamont, 20R15: p. 54) The tableau contains two candidates and four constraints. The last tIwBo constraints occupy the same hierarchical rank, hence, the dotted lines separating thLe third and fourth rows of the constraints. The first and most highly ranked constraiNnt prohibits the occurrence of a fricative as the onset of a syllable and the presence oAf voiceless dental fricative /θ/ in the onset of the first candidate fatally violates this Dconstraint. This candidate also violates constraint 2 which prohibits the occurrence ofA a segment that is fricative. The other two constraints prohibit any cIhBange in the place or manner of articulation of the correspondent segments in the inpFut a nd output phases. Thus, the substitution of voiceless dental fricative /θ/ with voiceless alveolar plosive /t/ has changed the place and manner of articulation of the segment f roOm dental and fricative to alveolar and plosive, respectively. In this way, candidate (bY) has violated constraints 3 and 4. T However, thesSe coInstraints are not as highly ranked as the first two constraints in the tableau, thRerefore, candidate (b)’s violation of them is not fatal and it still emerges as the optimaEl candidate in the tableau. GIeVnerally and across the data, it is noticed that the dental fricatives /θ ð/ are substituted Nwith alveolar plosives /t d/ respectively. The instance found in the longitudinal data above U complements this finding too. 122 4.3.2.6 Representative acoustic analysis of substitution There are two segment substitutions in the data and these concern the voiced and voiceless dental fricative (/ð θ/) with voiced and voiceless alveolar plosives (/d t/), respectively. One acoustic illustration for each of these substitutions has been carried out in the subsequent paragraphs; with and the have been used for this purpose. Y AR R LI B AN D IB A OF Y SI T R IV E UN 123 RY RA LI B AN BA D Fig. 4.4 Acoustic illustration of s IFubs titution in /wit/ “with” by a five-year-old O Y TSI R IV E UN 124 The segment under analysis in the acoustic illustration above is a consonant sound, specifically, a voiceless alveolar plosive /t/ which now occupies the position of coda in the analysed structure above. The acoustic features of this sound are highlighted in this analysis. The waveforms and the spectrograms are clearly separated into two different portions and these are the onset and peak (/wi/) in one portion, and the coda (/t/) in the other portion. There is evidence that the first group of spectrogram and waveforms represent Y/wi/ because bilabial semivowel /w/ shares some acoustic features with vowels anRd these features are represented by pseudo formants in acoustic analysis. Therefore, thAe absence of other formants for this first sound proves that it only has some semIbBlanc Re to a vowel, but is not really a vowel phoneme. L Thus, the presence of a first formant of about 300Hz fro m 1.27 seconds to about 1.54seconds is a representation of the first two sounds in thNis structure. There is an increase of the second formant from about 500Hz to 2555Hz tAowards 1.27seconds and this sudden increase in the second formant is an indicationA thaDt the articulation of that sound is close to the lips – it is a front vowel. Also, the valuBe of F1 at about 300 to 400Hz shows that the lips are more closed than open. Thus, the v oIwel that fits into these acoustic descriptions is /i/. Finally, there is a portion of OsilenFce between the first visible groups of spectrogram and waveforms and the second. T his portion of silence is an acoustic feature of plosives, which typically have a brieTf paYuse before an outburst of acoustic energy. This portion of silence is represented by Ithe white gap between the two groups. The dynamic range of the spectrograms wSas reduced to 40.0 to eradicate, as much as possible, any representation of backgrounRd noise. Thus, the portions of silence are more clearly represented by an absence of greyE shades. This portion of silence is followed by a friction noise whose outburst is not soI Vdistinct because it is a voiceless sound and because it is word final, hence, there is a Nfading quality to this phoneme. This instance of substitution was common to many of the U participants. However, one five-year-old male participant’s utterance was used as a representation of the others. Another instance of substitution is presented below, however, this second instance involves voiced dental fricative (/ð/) being substituted with voiced alveolar plosive (/d/). 125 RY RA LI B N DA Fig. 4.5 Acoustic illustration of substitutionA in /di/ “the” by a five-year-old IB F OY IT ER S NI V U 126 The centre of analysis in Figure 4.5 is a consonant phoneme, specifically, the substitution of voiced dental fricative /ð/ with voiced alveolar plosive /d/. The presence of the substitute in the acoustic analysis above is represented by the outburst of acoustic energy which extends from 100Hz to about 5000Hz; also, there is a clear display of the outburst as shown by the straight line of spectrogram that the dark shades of grey represent. This shows that the sound produced at that point is both plosive and voiced. This phoneme is cloYsely followed by clearly defined fundamental frequencies, showing that the segment foRllowing the voiced plosive is a vowel. This vowel phoneme has a high F2 value (1613H Az), showing that the articulation is close to the lips, and in this case, the phonemIe Bis /i/ R. 4.3.3 Constraints ranking for monophthongisation L Four diphthongs were tested for the process of monophthonNgisation among the participants. The tested diphthongs were /ai/, /ei/, /ǝu/, and /eǝ/. I, Avine, and time were used to test /ai/; a, gave, day, play, cake, take, rain, gate, later, anDd away were used to test /ei/; home, and Rose were used to test /ǝu/; and, there, hair anAd where were used to test /eǝ/. 4.3.3.1 Realisation of I IB This first person singular pronOounF, I, had two variants in the data and these are /ai/ (92%) and /a/ (8%). These variants are used as the input for Optimality analysis for the emergence of /ai/ in Tableau 4.8T. Y I RS E V UN I 127 Tableau 4.8 Emergence of /ai/ “I” I /ai/ MAX NODIPHTHONG a. /ai/ * RY b. /a/ *! RA IB L DA N A IB OF Y T RS I E NI V U 128 Constraints ranking: MAX >> NODIPHTHONG MAX: No deletion (Every phoneme in the base should have a correspondent in the output) (Wheeler, 2007: p. 3) NODIPHTHONG: diphthongs are prohibited (Green, 2010: p. 101) Tableau 4.8 contains two candidates and two constraints. Candidate (b) fatally violYates constraint 1 because of the absence of /i/ since the constraint prohibits any deletRion of a segment between the base and the output phases. The second violation in the Atableau is by the optimal candidate, candidate (a), this violation occurs because candidRate (a) remains a diphthong, whereas, diphthongs are prohibited by constraint 2. ThIiBs violation however, does not prevent the emergence of candidate (a) as the harm oLnic candidate because the constraint is not as highly ranked as constraint 1. The futiNlity of this violation is indicated by the grey shade of the cells under constraint 2, toA show that the violations under this constraint do not prevent the emergence of candidDate (a). 4.3.3.2 Realisation of gave BA Gave is one of the structures used to teIst the diphthong /ei/ among the participants. This word occurred once in the instrumFent and there were three variants of it as found in the data. These variants are /g ɛvO/ (94%), /geiv/ (5%) and /gav/ (1%). These variants are considered as the candidYates in the input for Optimality analysis that will account for the emergence of /gɛv/I iTn Tableau 4.9. S ER V UN I 129 Tableau 4.9 Emergence of /gɛv/ “gave” Gave /geɪv/ NODIPHTHONG IDENT-IO(VOICE) MAX a. /geɪv/ *! RY b. /gɛv/ * A c. /gɛf/ * R* LI B N AD A IB O F TY RS I IV E UN 130 Constraints ranking: NODIPHTHONG >> IDENT-IO(VOICE) >> MAX NODIPHTHONG: diphthongs are prohibited (Green, 2010: p. 101) MAX: No deletion (Every phoneme in the base should have a correspondent in the output) (Wheeler, 2007: p. 3) IDENT-IO(VOICE): Assign one violation mark for every voice feature in the input that is different from its corresponding feature in the output – underlying specifications for vYoice should be respected (Marc van Oostendorp, 2011: p. 3) AR NODIPHTHONG is the higher ranked constraint of the three constraints in tRhis tableau and it has been fatally violated by candidate (a) because of its retention oIf Bits closing diphthong /ei/. Candidates (b) and (c) however, prevent a violation of this hLigher ranked constraint by violating the lesser ranked constraint instead. In other words, ra ther than retain their closing diphthong /ei/, candidates (b) and (c) delete one of their conNstituent vowels in order to obey this constraint. The second constraint, IDENT-IOD(VOICAE), prohibits a change in the voice feature between the input and output phases, the change in the voice feature of voiced labiodental fricative /v/ to voiceless labiBodenAtal fricative /f/ in candidate (c) violates this constraint. Therefore, candidate (c) h asI violated the next highly ranked constraint in the tableau. This violation automaticFally leads to the emergence of candidate (b) as the harmonic candidate. Its emerOgence is not deterred by its violation of the third constraint which is the least rankedY con straint in the tableau. 4.3.3.3 RealisationI oTf rose Rose is onRe ofS the linguistic items used to test the closing diphthong /ǝu/ among the participEants. Rose occurred six times in the instrument and there were two variants of the parVticipants’ production of it. These variants are /roz/ (96%) and /rǝuz/ (4%). These variants aIre used as the candidates in the input for Optimality analysis that will account for the Nemergence of /roz/ in Tableau 4.10. U 131 Tableau 4.10 Emergence of /roz/ “rose” Rose /rǝuz/ NODIPHTHONG MAX Y a. /rǝuz/ *! R b. /roz/ * RA B L I AN BA D F I O TY I ER S V UN I 132 Constraints ranking: NODIPHTHONG >> MAX NODIPHTHONG: diphthongs are prohibited (Green, 2010: p. 101) MAX: No deletion (Every phoneme in the base should have a correspondent in the output) (Wheeler, 2007: p. 3) Tableau 4.10 has two constraints and two candidates. The first candidate (/rǝuz/) violYates the first and higher ranked constraint which prohibits the occurrence of a diphthoRng. This violation is fatal, and since there are only two candidates in the tableau, Acandidate 2 becomes the likely candidate to emerge as the optimal candidate. This emeRrgence is further made possible because candidate 2 has eliminated its diphthong andI reBplaced it with a pure vowel, thus, violating constraint 2 instead of constraint 1. ConLstraint 2 is not as highly ranked as constraint 1, so this violation is not fatal. Thus,N cand idate (b), /roz/, emerges as the optimal candidate. A 4.3.3.4 Realisation of where D The fourth diphthong that was tested is tBhe cAentring diphthong /eǝ/ and where, there and hair were used to test it. Where occ uIrred once in the instrument and the participants produced two variants for this strFucture. These variants are /we/ (97%) and /weǝ/ (3%). These variants are subjected tOo Optimality theoretical analysis in Tableau 4.11. Y IT RS VE UN I 133 Tableau 4.11 Emergence of /we/ “where” Where /weǝ/ NODIPHTHONG MAX Y a. /we/ * AR b. /weǝ/ *! BR I N L A BA D F I Y O T RS I E NI V U 134 Constraints ranking: NODIPHTHONG >> MAX NODIPHTHONG: diphthongs are prohibited (Green, 2010: p. 101) MAX: No deletion (Every phoneme in the base should have a correspondent in the output) (Wheeler, 2007: p. 3) There are two candidates and two constraints in Tableau 4.11. The first and higher ranYked constraint prohibits diphthongs while the second constraint prohibits deletion. To oRbey the higher ranked constraint, candidate (a) removes one of the vowels in itsA diphthong, violating constraint 2 rather than constraint 1. Candidate (b) however retaRins its diphthong and fatally violates constraint 1 as result. Thus, candidate (a) emIeBrges as the optimal candidate. L 4.3.3.5 Realisation of may AN This instance of monophthongisation is culled fromD the longitudinal section of the data and it involves a reduction of the diphthong /ei/ tAo /ɛ/. This change is very similar to the ones that were recorded in the cross-sectionalI dBata and the same constraints ranking apply to this instance too. The structure has beeFn th eoretically analysed in the tableau. O Y IT RS IV E UN 135 Tableau 4.12 Emergence of /mɛ/ “may” Where /weǝ/ NODIPHTHONG MAX Y a. /we/ * AR b. /weǝ/ *! R IB N L A BA D F I Y O T RS I E NI V U 136 Constraints ranking: NODIPHTHONG >> MAX NODIPHTHONG: diphthongs are prohibited (Green, 2010: p. 101) MAX: No deletion (Every phoneme in the base should have a correspondent in the output) (Wheeler, 2007: p. 3) There are two candidates and two constraints in Tableau 4.12 above. The first and higYher ranked constraint prohibits diphthongs while the second constraint prohibits deleRtion. To obey the higher ranked constraint, candidate (a) removes one of the vowels in itAs diphthong, violating constraint 2 rather than constraint 1. Candidate (b) however retaRins its diphthong and fatally violates constraint 1 as result. Thus, candidate (a) emIerBges as the harmonic candidate. L 4.3.3.6 Representative acoustic analysis of monophAthonNgisation There are five words that were used to test the procDess of monophthongisation, four of these linguistic items exhibited the process while Aone remained as a diphthong. Of these four instances, rose has been acoustically illuIsBtrated to show monophthongisation. F O TY RS I IV E UN 137 AR Y R IB N L AD A B Fig. 4.6 Acoustic analysis of monoph tIhongisation in /roz/ “rose” by a five-year-old OF Y IT S ER NI V U 138 Figure 4.6 above presents the acoustic analysis of rose. The first and second formants of the vowels in the word are the most prominent in the spectrogram and this is shown by the red dots that consistently line the central portion of the spectrogram where the peak, which is the vowel segment, is located. A closer observation of these formant dots shows that there is a slight descent in F2 values from the beginning of the peak to its end. The average values for F1 and F2 have been automatically retrieved from PRAAT and plotted onY the acoustic vowel chart below. F1 value is 467.31Hz and F2 value is 1049.78Hz. AR R IBL DA N A F IB O TY SI VE R I UN 139 F2 [Hz] Y 200 R 250 A 300 R 350 400 B I 450 o L 500 550 N 600 A 650 D 700 A 750 B 800 I F1 [Hz] Fig. 4.7 Acoustic vowel chart forF /o/ by a five-year-old O TY I S ER NI V U 140 2800 2700 2600 2500 2400 2300 2200 2100 2000 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 900 800 The position of this vowel on the acoustic chart above shows that its place of articulation on the tongue is at the back, and the height of the tongue during its articulation is somewhere between half close and half open. This vowel phoneme is much closer to half- close back rounded vowel /u/ in the closing diphthong found in the British English transcription of rose /rǝʊz/, than it is to half-open central neutral vowel /ə/. The slight descent of the second formant value which was noticed in the spectrogram is an indicaYtion that the articulation of this vowel segment is towards /u/, because F2 values are aAn inRdication of articulation distance from the lips, that is, towards the back of the tongue. The F2 value of this sound begins at 1283.925329Hz and ends at 991.499879Hz,B a clRear reduction of value. Although only one participant’s utterance was usedL toI derive this acoustic representation, the production is very similar to the majority of the participants in the data. 4.3.4 Constraints ranking for Assimilation AN Four linguistic structures were used to test the procDess of assimilation and these are of blue, like blue, Jane’s home and like this. Of bluAe had three variants; Jane’s home had two variants; like this had three variants a nIdB like blue had three variants in the data. All the instances of assimilation were regFressive in direction. Like blue is analysed in the tableau below and its variants serve as the candidates in the tableau. Also, two instances of the process of assimilation ha vOe been culled from the longitudinal data and analysed theoretically. These TstruYctures are have password and give me from the older and younger participants, respecItively. 4.3.4.1 ReaRlisaStion of like blue Like blEue occurred once in the instrument and there were three variants of it which are /lIaVigblu/ (95%), /laikblu/ (3%) and /lakblu/ (2%). These variants are used as the input for NOptimality analysis for the emergence of /laigblu/ in Tableau 4.13. U 141 Tableau 4.13 Emergence of /laigblu/ “Like blue” like blue /laik blu/ AGREE(voice) FIN VOI αF a. /laikblu/ *! *  b. /laigblu/ * RY c. /lakblu/ *! * *R A IB L AN BA D I O F TY RS I VE UN I 142 Constraints ranking: AGREE(voice) >> FIN VOI >> αF AGREE(voice): Assign a violation mark for every pair of adjacent consonants whose members differ in their specification for voice (Lamont, 2015: p. 54) FINAL VOICING: Phonemes in syllable coda must be voiced (Lamont, 2015: p. 55) αF: correspondent segments in input and output must have identical values (MustafaYwi, 2006: p. 76) R The tableau has three candidates and three constraints. There are two fatal violAations of the most highly ranked constraint by candidates (a) and (c). This constraRint prohibits the occurrence of contiguous consonant phonemes if they are differenIt Bin voice feature, and since candidate (a) has the voiceless alveolar plosive (/k/) o cLcurring beside the voiced bilabial plosive (/b/), it automatically violates this constNraint. Candidate (c) commits a violation in the same way too. A The second constraint (FIN VOI) which compels fiDnal phonemes to be voiced is violated by candidates (a) and (c) because their first syllaAbles end in voiceless velar plosives (/k/). The second candidate ends in a voiced ve laIrB plosive /g/ and does not violate this constraint. Constraint 3 requires faithfulness of the features of the output to that of the input, that is, the features in the output should mFatch those of the input. Candidates (b) and (c) violate this constraint because of th e Ochanges that have occurred between the input and the output stages. From these vTiolaYtions, candidate (b) emerges as the harmonic candidate because candidates (a) and I(c) already incurred fatal violations against constraint 1. 4.3.4.2 ReaRlisaStion of have password This inEstantiation of assimilation is also regressive and it affects the feature of voice bIecVause voiced labiodental fricative /v/ changes to voiceless /f/ as shown in the tableau. NThis change is engendered by the segment’s occurrence before voiceless bilabial plosive U /p/. This assimilation is theoretically analysed in the tableau. 143 Tableau 4.14 Emergence of /haf/ “Have” Have /hav/ σCONT *SCHWA MAX IDENT- IO(voice) a. /hav/ *! Y b. /haf/ * AR c. /əv/ * * BR L I DA N A IB OF Y T RS I E NI V U 144 Constraints ranking: σCONT >> *SCHWA >> MAX >> IDENT-IO(voice) σCONT: the onset of a syllable must not be of greater sonority than the last segment (Pons, 2005: p. 13) *SCHWA: The realisation of a central vowel is banned (Wheeler, 2007: p. 8) MAX: No deletion (Every phoneme in the base should have a correspondent in the outYput) (Wheeler, 2007: p. 3) IDENT-IO(voice): Every voice feature in the input that is different from its corresRponding feature in the output is prohibited – underlying specifications for voice shoRuld Abe respected (Marc van Oostendorp, 2011: p. 3) IB MAX and IDENT-IO(voice) are faithfulness constraints and they arLe more lowly ranked than the markedness constraints in the tableau; this is a mNajor observable feature in the participants’ utterances. Thus, this proves that theA most important thing is not just articulating towards the target pronunciation of thDe adult, rather, it is producing utterances as simplified as possible while trying to produAce the target language. The assimilation process in this inst anIcBe affects voiced labiodental fricative /v/ which occurs before voiceless bilabial plFosive /p/, although both occur in two different syllables. During production, the voiceless state of /p/ regressively influences /v/, hence assimilating it into the voiceless form /fO/. The emergence of candidate (b) (/haf/) as the optimal candidate means that theY other candidates have violated at least one of constraints σCONT, *SCHWA or MAX. I*TSCHWA prohibits the occurrence of a central vowel – this means that candidate (c) haSs violated this constraint because it contains half-open central neutral vowel /ə/. MVrE R IAX equires that the segments in the input are all present in the eventual output – this is Nnot the case with candidate (c) since the glottal fricative /h/ has been deleted already. U σCONT mandates that the onset of a syllable should not be of greater sonority than the last segment in it; this constraint is also violated by candidate (a) because /h/ is more sonorant than voiced labiodental fricative /v/. This violation is a fatal one because σCONT is the highest ranked constraint in Tableau 4.14. Also, the violation of *SCHWA by candidate (c) means that candidate (b) emerges as the optimal candidate. Proximity is a very important factor in local assimilation of segments to one another. Hence, although /v/ and /p/ belong 145 to different morphological units, the pause between them is very infinitesimal, almost non- existent, which prompts the regressive assimilation to occur quite easily and noticeably. 4.3.4.3 Realisation of give me This final instance of assimilation was also regressive and is got from the longitudinal data, particularly from the younger participant who represents the outset of child languYage acquisition. There were four instances of this mode of assimilation in this sectioRn of the data and these are give me which is produced as /'mimi/, packer as /'kaka/A, banana as /na'nana/ and water as /'tɔta/. Give me is theoretically analysed in the tabRleau below. B I L N AD A F IB Y O IT ER S IV UN 146 Tableau 4.15 Emergence of /mimi/ /gɪv mi:/ (give me) *DIFF[SON] MAX Y a. /gɪv mi:/ *! R b. /gɪf mi:/ *! A R c. /gimi/ *! * IB d. /mimi/ * L AN BA D IF O IT Y RS IV E UN 147 Constraints ranking: *DIFF[SON] >>MAX *DIFF[SON]: No sonority differences between onset consonants (Pons, 2005: p. 13) MAX: No deletion (Every phoneme in the base should have a correspondent in the output) (Wheeler, 2007: p. 3) There are four candidates and two constraints in the tableau above. The first constraYint, *DIFF[SON], prohibits a difference in the sonority of syllable onsets, and candidatRes (a) to (c) have violated this constraint because nasals are much more sonorous thAan plosives. Thus, the presence of voiced velar plosive /g/ in the onset of the first Rsyllable and the bilabial nasal /m/ in the onset of the second syllable in candidates (Ia)B to (c) is equal to the violation of this higher ranked constraint. Therefore, these vio laLtions are fatal. Candidate (d) however, has replaced /g/ with bilabial nasal /m/, tNhus preventing a violation and emerging as the harmonic candidate in the tableau. A Although candidates (c) and (d) violate constrainDt 2, these violations are trivial because MAX is not as highly ranked as the first consAtraint. Therefore, candidate (d) still emerges as the optimal candidate in the tableau . IB This participant was noticed tOo favFour the onsets of monosyllabic words over the codas so that monosyllabic words w ere articulated without their codas. This assimilation is a regressive one because iYts influence is backwards in direction. Also, unlike the assimilation process which is kInoTwn to affect voice, place or manner of articulation, this assimilation forces all of itsS features on the concerned phoneme; hence, it replaces the sound totally. In other wordRs, two processes simultaneously occur in these instances – assimilation and substitEution. The substitution does not happen arbitrarily, rather, a second onset replaces thIeV first. UN4.3.4.4 Representative acoustic analysis of assimilation Four structures in the instrument were used to test the process of assimilation among the participants, and of these four structures, like blue /laigblu/ has been acoustically analysed in the figure. 148 AR Y BR L I N AD A IB Fig. 4.8 Acoustic illustrationO of aFssimilation in /laigblu/ by a five-year-old TY SI VE R I UN 149 The particular segment under analysis in the figure above is the coda of the first syllable (voiceless velar plosive /k/) which has been assimilated to voiced velar plosive /g/. This segment is very subtly produced and takes only a small fraction (0.04 seconds) of the total milliseconds (0.89 seconds) used to produce the structure. Also, the portion of silence that is customary with plosives is absent from this segment, and the only concentration of acoustic energy that shows that it is a voiced sound is at the base of the fundameYntal frequency where dark shades are visible. This portion of the syllable is very subAtle aRnd even though only one participant’s utterance has been used to generate this acoustic analysis, it can be regarded as representative of a majority of the participants. IB R 4.3.5 Constraints ranking for deletion L Nine linguistic structures were used to test the process of dNeletion and these were next, set, grandma, grandpa, and, friends, combed, want, Aand textbooks. Deletion occurred significantly in all the linguistic structures excepDt friends where many of the participants maintained the number of the consonant phonAemes in the margins. Thus, textbooks which had a significant portion of deletion andI frBiends which had limited deletion are theoretically analysed in the subsequent tableaux. Textbooks and friends occurred once each in the instrument and have three andO fourF variants, respectively. 4.3.5.1 Realisation of TYextb ooks The variants of the IprToduction of textbooks are /tezbuk/ (68%), /tezbuks/ (15%), /tekstbuks/ (11%), and /tekSsbuk/ (6%). These variants serve as candidates in the input for Tableau 4.16. ER NI V U 150 Tableau 4.16 Emergence of /tezbuk/ “textbooks” Textbooks /tekstbuks/ *COMPLEX MAX a. /tekstbuks/ **!! Y b. /teksbuks/ **!! * AR c. /tesbuks/ *! * R d. /tezbuks/ *! LI* B e. /tezbuk/ N * A D IB A O F Y IT ER S IV UN 151 Constraints ranking: *COMPLEX >> MAX *COMPLEX: a syllable has no more than one consonant in a margin (Wheeler, 2007: p. 3) MAX: No deletion (Every phoneme in the base should have a correspondent in the output) (Wheeler, 2007: p. 3) Tableau 4.16 contains five candidates and two constraints. All the violations agaYinst constraint 1 are fatal because it is the higher ranked constraint. Candidates (a) Rand (b) violate this constraint twice each, hence the double asterisks and exclamationA marks. This constraint prohibits complex margins, that is, a margin – whether onseRt or coda, is not permitted to have more than one consonant phoneme. The preseIncBe of more than one phoneme in the coda of candidates (c) and (d), and the presence oLf more than one consonant phoneme in the coda of the first and second syllables of canNdidates (a) and (b) are violations of this constraint. Only candidate (e) obeys this conAstraint by getting rid of its excess consonant phonemes. Thus, candidate (e) emergesD as the optimal candidate. This emergence is still valid despite candidateA (e)’s violation of the second constraint which prohibits any form of deletion. This is bIecBause the second constraint is not as highly ranked as the first which it has obeyed. ThFe se cond constraint (MAX) is also violated by four other candidates, (b) to (e), since soOme segments have been deleted. Because the violations uYnde r constraint 1 have already produced an optimal candidate, the rest of the tableauI iTs shaded in grey colour, showing the limited relevance of whatever happens there tSo the emergence of /tezbuk/. 4.3.5.2 ReaRlisation of Friends ThVe vaEriants of the enunciation of friends were /frendz/ (58%), /frens/ (29%), and /frend/ (1I3%). These variants serve as candidates in the input for Tableau 4.17. UN 152 Tableau 4.17 Emergence of /frend/ Friends /frendz/ MAX *COMPLEX a. /frend/ *! ** Y b. /frendz/ ** AR c. /frens/ *! ** R LI B N AD A IBF Y O SI T VE R I UN 153 Constraints ranking: *COMPLEX>> MAX *COMPLEX: a syllable has no more than one consonant in a margin (Wheeler, 2007: p. 3) MAX: No deletion (Every phoneme in the base should have a correspondent in the output) (Wheeler, 2007: p. 3) There are three candidates and two constraints in Tableau 4.17. The higher ranYked constraint (MAX) prohibits the deletion of any phoneme, but only candidate (bA) obReys this constraint, while candidates (a) and (c) violate it through a deletion of voiced alveolar fricative /z/ and voiced alveolar plosive /d/, respectively. Thus, the Rmajority of the participants retained the consonant clusters in this linguistic strucItuBre, unlike textbooks which was analysed in the previous tableau. The motivation fo r tLhis variation is difficult to decipher because both structures are made up of almost thNe same types of phonemes. 4.3.5.3 Realisation of want A This instance of deletion is derived from the AlongDitudinal data, from the older participant specifically. This structure is analysed iIn Bthe next tableau. OF Y IT ER S V UN I 154 Tableau 4.18 Emergence of /wɔn/ want Want /wɔnt/ *COMPLEX MAX Y a. /wɔnt/ *! AR b. /wɔn/ * BR I N L A BA D F I Y O T RS I E NI V U 155 Constraints ranking: *COMPLEX >> MAX MAX: No deletion (Every phoneme in the base should have a correspondent in the output) (Wheeler, 2007: p. 3) *COMPLEX: a syllable has no more than one consonant in a margin (Wheeler, 2007: p. 3) The structure analysed in Tableau 4.18 is taken from a string of utterances and the featuYres of context have inadvertently engendered the phonological process of deletionR in the analysed lexical structure. This structure /wɔnt/ ends in the voiceless alveolar Aplosive (/t/), this is the phoneme which also began the next word in the data. Thus,R this presents an instance of juncture prosody in continuous speech which leads to IthBe elision of the first voiceless alveolar plosive. As a result of this, the deletion of / t/ Lis encouraged, especially, when it has occurred in-between two consonant phonemeNs, the second of which is also a voiceless alveolar plosive /t/. A Considering the first constraint in the tableau abovDe, the retention of a complex coda in the first candidate leads to a fatal violation of theA higher ranked constraint, leaving the second candidate to emerge as the optimal cIanBdidate. The activities under this constraint are enough to determine which of the tw o candidates is harmonic; thus, the other row in the tableau is shaded grey to showO thatF the activities in them do not affect the overall emergence of the already decided harm onic candidate under constraint 1. The second and lowIeTr ra Ynked constraint (MAX) in the tableau is violated by candidate (b) but this does noSt engender any change because it is not as highly ranked as the first. Also, candidate (Rb)’s obedience of constraint 1, the higher ranked constraint, already leads to its emergeEnce as the harmonic candidate. 4I.3V.5.4 Realisation of come NThis is another instance of deletion as found in the longitudinal data, however, this is U retrieved from the younger participant’s data at the outset of the longitudinal data collection at one year, three months. This participant constantly elided the final consonants of monosyllabic words so that words like come, take, phone, food, ball and fine were enunciated as /kͻ:, tɛ, fo, fu, bͻ:, fa/, respectively. “Come” /ko/ has been analysed below. 156 Tableau 4.19 Emergence of /kɔ/ come Come /kʌm/ *C]σ MAX  a. kɔ: * Y b. kɔm *! AR c. kʌm *! R LI B N AD A IBF Y O IT ER S IV UN 157 Constraints ranking: *C]σ >> MAX *C]σ - Syllables must not have codas. (Borowsky, 1989: p. 146) MAX: No deletion (Every phoneme in the base should have a correspondent in the output) (Wheeler, 2007: p. 3) The first constraint, *C]σ, is a markedness constraint which prohibits syllables from havYing codas. This constraint ranks higher than MAX and its violation by candidates (b)R and (c) leads to the emergence of candidate (a) as the optimal candidate. Since *C]σ Aranks higher than MAX, candidate (a)’s violation of MAX is not relevant because EVALR has determined the harmonic candidate. Hence, the second column of the constraints IisB shaded grey because no matter the violations that occur there, the emergence of the oLptimal candidate (/kɔ/) is not affected. N With the deletion of the final consonant in the analysAed word, the phonological grammar of this participant elevates the markedness consDtraint above the faithfulness constraint because the eventual output /kɔ/ is not ideBnticAal with the input of /kʌm/, a requirement that ranks highly in faithfulness constraints . IFor a child at the outset of the use of the expressive language skill, this production is sFimilar to that of the target production and intelligibility is still achieved. It is noted thOough, that whenever the participant used this word, it was accompanied with a physYical gesture of the hand which must have been imitated from adults around. This gives mTore credence to the contribution of imitation in a child’s language acquisition processI than the literature accords it. This participant began to produce this word at agRe 1;S7 and continued to use it beyond this time. Comprehension of this word alreadyE manifested before this time. 4I.3V.5.5 Representative acoustic illustration of deletion UNNine linguistic structures were used to test the process of deletion as realised among the participants. Of the nine structures, second /sekɔn/ has been acoustically analysed in the figure below. 158 AR Y BR L I N AD A IB F Fig. 4.9 Acoustic analysis o f Odeletion in “second” /'sekɔn/ by a four-year-old YIT ER S IV UN 159 The coda in the second syllable of this structure has been deleted. Meanwhile, the phonemes that are left can be identified by the spectrogram represented by the different shades of grey, and these begin from about 5.57seconds on the spectrogram. This portion of the analysis begins with a burst of acoustic energy which is representative of a plosive. This plosive is not a voiced one and this is evident by the absence of fundamental frequency which shows the inactivity of the vocal cords. This wave is followed by clear formants Ythat are indicative of a vowel segment. Vowels are always voiced and the indicationR for the voiced state, the fundamental frequency, extends into the last parts of the spectrAogram. This shows that a new sound ends this structure and this sound has the simBilitudRe of a vowel but is not a vowel, as the other formants are absent. In other words,L theIre is an indication of a nasal and nothing more – the plosive which ought to have ended the structure is absent from the spectrogram which means that it has been elided. AN 4.3.6 Constraints ranking for vowel strengtheDning The central vowels /ʌ ɜ: ǝ/ were tested for thAe process of vowel strengthening among the participants. Twelve words which contaIinBed these vowel phonemes were inserted into the instrument. These words are sun, wFate r, bud, cut, up, mum, us, our, month, little, away and birthday. Vowel strengtheningO occurred significantly in all the linguistic structures. Five of these words have beenY pur posively selected for theoretical analysis in the subsequent tableaux. Each centTral vowel is represented, and the other two words have peculiar structures which haIve prompted their representation in the tableaux. S 4.3.6.1 ReaRlisation of Cut CuVt waEs used to test vowel strengthening in open central neutral vowel /ʌ/. There were two vIariants during the participants’ pronunciation of cut, and these are /kɔ:t/ (97%) and /kʌt/ N(3%). These variants are presented as candidates in the Optimality analysis in Tableau 4.20. U 160 Tableau 4.20 Emergence of /kɔt/ “cut” Cut /kʌt/ *SCHWA αF a. /kʌt/ *! RY b. /kɔ:t/ * RA B I L AN D IB A O F TY RS I E NI V U 161 Constraints ranking: *SCHWA >> αF *SCHWA: The realisation of a central vowel is banned (Wheeler, 2007: p. 8) αF: correspondent segments in input and output have identical values for [voice], [high] and [cont] (Mustafawi, 2006: p. 76) There are two constraints and two candidates in the tableau above. *SCHWA is the higYher ranked constraint and it forbids the realisation of a central vowel. This constraiRnt ranks higher than the second constraint, αF, which is a faithfulness constRraintA and which encourages input and output segments to have identical values for eacBh feature. Candidate (a) violates *SCHWA because its peak is a central vowel. CIandidate (b) however, obeys this constraint because its peak is not a central vowel, bu t aL back vowel. By changing its peak from central to back vowel, candidate (b) violatNes the second constraint which mandates that corresponding segments in input and oAutput should have identical values. Candidate (b)’s violation of this second constraDint makes it to emerge as the optimal candidate because constraint 2 is not as highlyA ranked as constraint 1. The rest of the words that contain open cIenBtral neutral vowel /ʌ/ and were used to test vowel strengthening underwent this proceFss of vowel strengthening too so that sun, bud, up, mum, us, and month were produced Oas /sɔ:n/, /bɔ:d/, /ɔ:p/, /mɔ:m/, /ɔ:s/ and /mɔ:nt/, respectively. Like half-open central nYeutr al vowel /ǝ/, /ʌ/ is also a subtle sound and enjoys a peculiar pronunciation by NIiTgerian speakers of English. This sound is absent from many Nigerian languages whicSh have much fewer vowel phonemes than the English language. Naturally therefore, Rspeakers unconsciously replace these phonemes with similar ones which are present in Nigerian languages. Thus, the constraint *SCHWA, a markedness constraint that forVbidsE the subtle central vowels, is ranked higher than the faithfulness constraint which foIrbids any changes to be made to the features in the input. UN4.3.6.2 Realisation of Our The triphthong, our, was used to test half-open central neutral vowel /ǝ/ and it occurred five times in the instrument. There were two variants for this word which are /awa/ (80%) and /auǝ/ (20%). /ǝ/ was tested in this word and also analysed in Tableau 4.21. 162 Tableau 4.21 Emergence of /awa/ “our” Our /aʊǝ/ *COMPLEXVOW *SCHWA αF Y a. aʊǝ *! * AR  b. awa R* B c. awǝ * L I * DA N IB A O F IT Y S VE R UN I 163 Constraints ranking: *C VOW OMPLEX >> *SCHWA >> αF *C VOWOMPLEX : No string of vowels within a syllable (Green, 2010: p. 101) *SCHWA: The realisation of a central vowel is banned (Wheeler, 2007: p. 8) αF: correspondent segments in input and output must have identical values (Mustafawi, 2006: p. 76) Y There are three candidates and three constraints in the previous tableau. The inputR itself is a triphthong, hence, it has no consonant phoneme at all. Of the three generatedA candidates, only one retains the half-close back rounded vowel /u/ while the other twRo have the fully glided version of /u/ which is the bilabial semivowel /w/ in themI.B The first constraint prohibits the occurrence of triphthongs within a syllable; hence ,L candidate (a) violates this constraint by retaining its triphthong. Candidates (b) andN (c) however have obeyed this constraint by converting /ʊ/ to /w/. This constraint (*CA VOWOMPLEX ) is the highest ranked in the tableau, so its violation by candidate (a)D is fatal. As for the second constraint, *SCHWA, it prohibits the presence of a centAral vowel in a syllable and its realisation in candidates (a) and (c) show a violation wBhich leads to the emergence of candidate (b) as the harmonic candidate. Although can diIdate (b) also violates the third constraint, αF, which mandates that correspondent segmFents in input and output must have identical values, its emergence as the optimal ca nOdidate still stands. 4.3.6.3 Realisation of AYway T This disyllabic worId away was also used to test vowel strengthening and it occurred twice in the instrRumeSnt. It has been purposively selected for theoretical analysis because it has schwa Eas the only phoneme in the first syllable. There were three variants and they are /awVei/ (95.5%), /awɛ/ 3.5% and /ǝwei/ 2%. Half-open central neutral vowel /ǝ/ is the central vIowel that was tested for this process. Tableau 4.22 presents the analysis for this UNemergence. 164 Tableau 4.22: Emergence of /awɛ/ “away” Away /ǝwei/ *SCHWA NODIPHTHONG αF Y a. /awei/ * * AR b. /awɛ/ BR* I c. /ǝwei/ *! * N L A BA D I OF IT Y RS IV E UN 165 Constraints ranking: *SCHWA >> NODIPHTHONG >> αF *SCHWA: The realisation of a central vowel is banned (Wheeler, 2007: p. 8) NODIPHTHONG: Diphthongs are prohibited (Green, 2010: p. 101) αF: Correspondent segments in input and output have identical values for [voice], [high] and [cont]. (Mustafawi, 2006: p. 76) Y The three variants from the data are the candidates in Tableau 4.22, and there aRre three constraints as well. The first and most highly ranked constraint (*SCHWA) pArohibits the realisation of any central vowel, so the presence of half-open central neuRtral vowel /ǝ/ in candidate (c) engenders a fatal violation as indicated by the asterisIk Band the exclamation mark. The second constraint also prohibits the presence of a dipLhthong and candidates (a) and (c) violate this constraint. The final constraint in theN tab leau prohibits any form of change in the values of a segment from the input to tAhe output phase. Candidates (a) and (b) violate this faithfulness constraint. Despite thDis violation however, candidate (b) still emerges as the optimal candidate because of tAhe violations that have been incurred against constraints 1 and 2 by the other candidIatBes. This is responsible for the shaded cells under constraint 3, because regardless of the violations that all the candidates incur, candidate (b) will still emerge as the harmonic cFandidate. 4.3.6.4 Realisation of birthd aOy Birthday occurred tIhTree Ytimes in the instrument, and has in fact been used to test the process of substitution.S However, this word also contains half-close central neutral vowel /ɜ:/, so it was also uRsed to test the process of vowel strengthening. There were four variants of the participEants’ production and these are /bedɛ/ (86%), /bɜ:θdɜ/ (6.7%), /betdɛ/ (6.3%) and /bɜV:θdei/ (1%). The variants are the candidates in Tableau 4.23 below. However, it is nIecessary to reformulate these frequencies since vowel strengthening is what is being UNconsidered here, and this occurred in more than one variant which are /bedɛ/ (86%) and /betdɛ/ (6.3%); thus, it can be logically concluded that vowel strengthening occurred in 92.3% of the total of variants. For the sake of the analysis in the tableau, the four variants are inserted as candidates. 166 Tableau 4.23: Emergence of /bedɛ/ “birthday” Birthday /bɜ:θdei/ *SCHWA NODIPHTHONG *FRICATIVE αF Y a. /bedɛ/ A*R b. /bɜ:θdɛ/ *! * R * IB c. /betdɛ/ *! L * d. /bɜ:θdei/ *! * AN * BA D I OF Y IT ER S NI V U 167 Constraints ranking: *SCHWA >> NODIPHTHONG >> *FRICATIVE >> αF *SCHWA: The realisation of a central vowel is banned (Wheeler, 2007: p. 8) NODIPHTHONG: diphthongs are prohibited *FRICATIVE: Segments may not be [+cont, –son] (Pater and Barlow, 2003: p. 492) αF: Correspondent segments in input and output have identical values for [voice], [high] and [cont] (Mustafawi, 2006: p. 76) RY There are four constraints in the tableau as well as four candidates. The firAst constraint prohibits central vowels and this constraint is violated by candidates (b) tRo (d) because of their retention of half-close central neutral vowel (/ɜ:/). Candidate (aI) hBowever replaces this central vowel with a half-open front spread vowel /e/, thus, avoLiding the violation of the most highly ranked constraint. This way, candidate (a) emNerg es as the optimal candidate. However, there are more violations in the tableau. CoAnstraint 2, NODIPHTHONG, prohibits the occurrence of a diphthong and this is violateDd by candidate (d) which still retains its diphthong. The third constraint is *FRICATIVAE and it prohibits the occurrence of fricatives in a structure. This constraint is also vIiBolated by candidates (b) and (d) because of the presence of voiceless dental fricative ( /θ/) in the codas of their first syllables. The final constraint is αF and it prFohibits any change in feature or value between the input and the output phases. Candi dOates (a) to (c) violate this constraint. This violation is the only one incurred by candidaYte (a) and it emerges as the optimal candidate in this tableau. 4.3.6.6 RealisationI oTf another S Another is Rculled from the longitudinal data and it contains two different central vowels – open cEentral neutral vowel /ʌ/ and half-open central neutral vowel /ǝ/ which occurs twice. TIhVese vowel phonemes have been strengthened. This instance of vowel strengthening Noccurred in the older participant’s utterance. Other instances of words from her that contain U this process are come, again, sister, around and open. Another has been analysed in the next tableau. 168 Tableau 4.24 Emergence of /anɔda/ “another” Another /ǝnɅðǝr/ *SCHWA αF *F-ONS Y a. /anɔda/ * AR b. /ǝnɅðǝr/ ***! R* LI B N AD A IBF Y O SI T VE R I UN 169 Constraints ranking: *SCHWA >> αF, *F-ONS *SCHWA: The realisation of a central vowel is banned (Wheeler, 2007: p. 8) αF: correspondent segments in input and output must have identical values (Mustafawi, 2006: p. 76) *F-ONS: A fricative cannot occupy the onset position (Pater and Barlow, 2003: p. 496)Y This instance of vowel strengthening was chosen for theoretical analysis because it Rcontains two of the three central vowels, one of which appears twice. Thus, the tableau Ais unlike the previous tableaux already engaged in this section. The uniqueness of tBhis tRableau lies in the fatal violations that occur under the first and highest ranked constIraint by candidate (b). This fatal violation occurs on three levels since there are t hLree central vowels in the candidate. The highly ranked constraint prohibits the occuNrrence of a central vowel, thus, violation is multiple. A The other two constraints occupy the same levelD of hierarchy and the first, αF, prohibits any change between output and input. The otAher constraint, *F-ONS, prohibits a fricative from occupying an onset position. B oIthB candidates violate either of these constraints. However, the emergence of an optimal candidate is quite clear because of the multiple violations that candidate (b) has inFcurred against *SCHWA, the highest ranked constraint. Candidate (a) on the other h aOnd, substitutes all three central vowels with non-central and stronger ones, thus, eTmeYrging as the optimal candidate. SI ER V NI U 170 4.3.6.7 Representative acoustic analysis of vowel strengthening RY RAB L I AN AD IBF Fig. 4.10 Acoustic analysis oOf vowel strengthening in “Cut” /kɔ:t/ by a four-year-old Y SI T R IV E UN 171 The replacement of the open central neutral vowel /ʌ/ with open back rounded vowel /ɔ:/ occurred frequently in the data and one of these instances has been examined acoustically. In the PRAAT analysis, the vowel portion of the utterance has been selected to capture the formants and the average first and second formant values were extracted. These were found to be 603.98Hz and 1202.08Hz for F1 and F2, respectively. These values were plottedY on the acoustic vowel chart below and the phoneme was found to occupy the half-opRen back position. RA LI B N AD A IBF Y O SI T VE R I UN 172 F2 [Hz] Y 200 R 250 A 300 R 350 400 B I 450 L 500 550 N ɔ: 600 650 DA 700 BA 750 800 I F1 [Hz] Fig. 4.11 Acoustic vowel chart ofF vo wel strengthening in /kɔ:t/ cut O IT Y S R IV E UN 173 2800 2700 2600 2500 2400 2300 2200 2100 2000 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 900 800 The figure above shows the position of the peak in cut as produced by the majority of the participants (97%). Although this vowel is plotted based on data got from the vocal tract resonances, its position is very similar to the articulatory chart which is based on the greatest point of tongue constriction only. Plotted using the first and second formant values, this segment /ɔ:/ fits the description of an open back vowel. Back and front vowels are acoustically stronger than central vowels, thus, the central vowel which occupied the pYeak of this structure has been strengthened and relocated to the back on the acoAustiRc vowel chart. R 4.3.7 Gliding IB The process of gliding involved the two semivowels /j/ and /w/ . ILt occurred in the word our which occurred five times in the instrument and in I am, NI ate and I have which occurred once each in the instrument also. In the case of oAur, there were two variants of the articulation by the participants and these are D/awa/ and /auǝ/ with 80% and 20% occurrence, respectively. Thus, there is an 80%A occurrence of gliding from /u/ to /w/ in this linguistic structure. This instance of gliIdiBng is theoretically analysed in the tableau below, and the variants constitute the candFida tes in the tableau also. Y O IT RS IV E UN 174 Tableau 4.25 Emergence of /awa/ Our /aʊǝ/ *COMPLEXVOW αF RY a. /aʊǝ/ *! RA  b. /awa/ I* B L AN BA D I O F TY RS I E NI V U 175 Constraints ranking: *C VOWOMPLEX >> * αF C VOWOMPLEX : No string of vowels within a syllable (Green, 2010: p. 101) αF: Correspondent segments in input and output have identical values for [voice], [high] and [cont]. (Mustafawi, 2006: p. 76) There are two candidates and two constraints in the tableau. The input itself is a triphthoYng, hence, it has no consonant phoneme at all. Of the two generated candidates, one dRoes not contain a consonant phoneme while the other contains a bilabial semivowel /Aw/. The first constraint prohibits the occurrence of triphthongs, hence, candidate R(a) violates this constraint by retaining its triphthong. Candidate (b) adapts itself to IobBey this constraint by converting /ʊ/ to bilabial semivowel /w/. The first constraint,L *C VOWOMPLEX , is more highly ranked, so candidate (a)’s violation of it is fatal. NAs f or the second constraint, it prohibits a difference in the values of an output fromA what obtained in the input. Thus, candidate (b)’s changes have violated constrainDt 2 which is not as highly ranked as constraint 1. Therefore, candidate (b) emergesA as the harmonic candidate. This instance of gliding was also fou ndI Bin the longitudinal data collected from the older participant. The commonness of tFhis instance to both groups, as well as the high rate of occurrence in the cross-sectional data confirms that complex vowels, especially triphthongs, are sourcesY of pr Oonunciation hiccups for many second language speakers of English in Nigeria. T 4.4 Processes eIxclusive to the longitudinal data 4.4.1 EpRenthSesis EpenthEesis occurred in the older participant’s speech in the longitudinal data and this inIvVolved an insertion of a close front spread vowel (/i/) between two contiguously occurring Nvoiced alveolar fricatives /z/. But for the intervention of this epenthetic vowel sound, the U only process that would have occurred in this utterance would have been a regressive assimilation of voice from /s/ to /z/ which would then lead to a gemination of /z/. This particular instance of epenthesis is analysed in the tableau below. Apart from this instance however, there is also epenthesis in this participant’s articulation of it, which was produced as /i'to/. 176 4.4.1.1 Epenthesis in “Press zero” Tableau 4.26 Emergence of /preziziro/ Y Press zero /preziziro/ *SS AGREE (VOICE) O-CORNTIGUITY a. /prez ziro/ *! RA  b. /preziziro/ IB * c. /pres ziro/ *! * L N A AD IB OF Y SI T R IV E UN 177 Constraints ranking: *SS >> AGREE(VOICE) >> O-CONTIGUITY *SS: Two sibilants are prohibited from occurring adjacently (Borowsky, 1989: p. 148) AGREE(VOICE): Assign one violation mark for every pair of adjacent consonants whose members differ in their specification for voice (Lamont, 2015: p. 54). O-CONTIGUITY: this constraint prohibits inserting any epenthetic phoneme between elements standing in correspondence so that there is no disruption in the adjaceYncy relationship in the output (Green, 2010: p. 103) R A Tableau 4.26 presents an analysis of epenthesis involving the insertionB of cRlose front spread vowel /i/ between two adjacently occurring alveolar fricatives /s/ aInd /z/. There are three candidates in the first row to be evaluated by EVAL. The fi rsLt and most highly ranked constraint (*SS) prohibits the occurrence of two sibilantNs adjacently and this is what is found in candidates (a) and (c), so the occurrence of /zAz/ and /sz/ in these candidates means that they have violated this constraint. This viDolation is a fatal one and leads to the emergence of candidate (b) as the harmonic cAandidate. The second constraint, AGREE(VOICE), p roIhBibits a difference in the specification for voice of a pair of adjacent phonemes. CandFidate (b) has two alveolar fricatives occurring together, however, both differ in theirO specification for voice. That is, while the first sound is voiceless, the second is voic ed and this is prohibited by constraint 2. Hence, constraint 2 has been violated bTy caYndidate (c) although this violation is not a fatal one. The final violation in the taIbleau is also by candidate (c), the optimal candidate. However, this violation is farS from being fatal because the rank of the constraint is the lowest in the tableau, soR its violation does not have a major impact on the emergence of the output as the harmonEic candidate. The constraint violated is O-CONTIGUITY and it prohibits the insertion oIf Van epenthetic phoneme between adjacent elements. This constraint is anti-epenthesis; Nhowever, its potential is greatly restricted because it does not occupy a high rank in the U tableau. In adult speech, it is not impossible to produce two sibilants together, however, in the utterance of the older participant, simplification is the most important underlying factor, hence contiguously occurring sibilants is not the first choice. 178 4.4.1.2 Representative acoustic analysis of epenthesis RY RA LI B N AD A IB OF Fig. 4.12 Acoustic analYysis of epenthesis in /preziziro/ by a four-year-old ITS R VE NI U 179 As a result of the epenthesis in the structure above, there are four peaks and are indicated by the upward arcs in the blue line that represents pitch in the acoustic analysis. The highlighted portion is the second peak and represents the epenthesised vowel in the structure. The formants are also clearly visible to authenticate that the area indeed represents the production of a vowel. Y 4.4.2 Devoicing R Certain segments within some linguistic structures were relieved of their voice Afeature even though they were produced in isolation of an influencing context or envRironment. These instances are classified under devoicing and were producedL mIai Bnly by the younger participant in the longitudinal data. The instances of devoici ng found in the data affect “dog” and “bag” (/dͻg, bæg/) which were produced as /dɔ:Nk, bak/, respectively. A AD IB OF IT Y S VE R I UN 180 Tableau 4.27 Emergence of /dͻk/ Dog /dͻg/ FIN DEV *VOICEDOBS]σ IDENT(voice) αF Y a. dͻg *! *! AR  b. dͻk * R* LI B AN BA D I F Y O SI T VE R I UN 181 Constraints ranking: FIN DEV, *VOICEDOBS]σ >> IDENT(voice) >> αF FINAL VOICING: Phonemes in syllable coda must be voiced (Lamont, 2015: p. 55) *VOICEDOBS]σ: Coda obstruents are devoiced (Mustafawi, 2006: p. 72) IDENT(voice): Voice feature remains the same in input and output (Oostendorp, 2011: p. 3) αF: This constraint compels correspondent segments in input and output to have identical values for [voice], [high] and [cont] (Mustafawi, 2006: p. 76). RY There are two candidates and four constraints in Tableau 4.27. The first twoA constraints jointly rank higher than the other two constraints. This is why the dominaRtion symbol (>>) does not appear between FIN DEV and *VOICEDOBS]σ. The doublIeB lines in the tableau between the third and the fourth rows are also an indication of Lthis demarcation between the constraints, while the dotted lines between constraintsN 1 a nd 2 show that they occupy the same rank. The first two constraints are responsibAle for the devoicing of voiced velar plosive /g/ to voiceless velar plosive /k/ in “dog”. DThe consonant in the coda of the input is /g/, and candidate (b)’s obedience of the first tAwo constraints means that the sound loses its voice feature by the time it emerges asI aBn optimal output. Hence, /dɔ:g/ becomes /dɔ:k/ after obeying the constraints. Voice is the only feature that changed, while the manner and place of articulation of the segmenFt remained the same. The other two constraints are not as highly ranked as the first tOwo; hence, candidate 2’s violation of these constraints does not change its emergencYe as the harmonic candidate. The highest ranking constraints in the tableau are markedIneTss constraints, ranking higher than the faithfulness constraints. This devoiced Sstate of the coda in /dɔ:k/ is another attempt from the participant at simplifEyingR segments. Beyond this explanation however, it is deduced from the data that theVre is a particular phonological context that prompts this devoicing. This context involves thIe appearance of a voiced consonant phoneme as the final phoneme in the coda of a Nsyllable. This theory is proposed because the participant was able to produce voiced U consonant phonemes when they appeared at the beginning of syllables and words where the phonological contexts are quite different from the consonants that have been devoiced. Participant produced “dog” /dɔ:k/ at nineteen months and “bag” /bak/ at twenty-one months old. 182 RY RA IB L AN D Fig. 4.13 Acoustic analysis of devoicingB in “Adog” /dɔ:k/ by a one-year-old I F O Y IT ER S NI V U 183 The final consonant in this structure has been devoiced and this is shown in the lack of a fundamental frequency (F0) in the highlighted portion of the spectrogram where the final segment is displayed. There is evidence that this sound is plosive because of the portion of silence that precedes it as well as the burst of energy shown in the upper half of the window. However, the lack of grey shades at the base of this analysis area shows that there was no glottal activity to make the sound voiced. Y 4.4.3. Stopping AR There are several instances of this in the younger participant’s utterancRes and these all involve the voiceless alveolar fricative /s/ being replaced by the voiIceBless alveolar plosive /t/. These occurrences occur in sit down and sit up, see (bag), a nLd see (biscuit) /ti, ti, ti/. AN AD IB O F Y SI T VE R NIU 184 Tableau 4.28 Emergence of /ti/ “Sit” Sit /sɪt/ *F-ONS AGREE(place) a. sɪt *! Y b. sɪ *! AR  c. tɪ * IB R LN AD A IBF Y O SI T R VE UN I 185 Constraints ranking: *F-ONS >> AGREE(place) *F-ONS: A fricative cannot occupy the onset position (Pater and Barlow, 2003: p. 496) AGREE(place): input and output must have identical place features (Lamont, 2015: p. 54) To account for the emergence of /ti/ as the optimal candidate in Tableau 4.28, *F-ONS is ranked higher and it prohibits a fricative from existing in an onset position. It Yis a markedness constraint. It prohibits the use of a fricative as an onset consonant. ThRis gives other consonant sounds, not just the voiceless alveolar plosive /t/ the opportuniAty to occupy the onset position. R Therefore, candidates (a) and (b) violate constraint 1 which is the hiIgBher ranked constraint in the tableau, leading to the emergence of candidate (c) as the o pLtimal candidate. Although candidate (c) violates constraint 2, this violation is no lonNger relevant because /ti/ already emerged as the optimal candidate. A It is important to note that all of the instances wheDre voiceless alveolar fricative /s/ occurs, the participant replaced them with voicelesAs alveolar plosive /t/; however, where the voiceless alveolar and voiceless palato-aIlBveolar fricatives /ʃ ʧ/ occurred, she replaced them with voiceless alveolar fricative /sF/. T he only consistency observed in this part of the data regarding these details is the replacement of a voiceless phoneme with other voiceless phonemes. The acousticY ana ly Osis of this instance is presented next. IT ER S NI V U 186 RY RA LI B N DA Fig. 4.14 Acoustic analysis of stopping Bin “SAit” /ti:/ by a one-year-old F I O TY SI VE R I UN 187 The general acoustic features of fricatives and plosives help to immediately realise that the sound on display in the spectrogram is not a fricative but a plosive. This is evidenced by the sudden burst of acoustic energy that begins the spectrogram. The vowel is also recognised by the visible first and second formants. Also, the distance of hertz between the first and second formants indicates that the vowel is a high one. Y 4.5 Summary R This chapter has presented the analysis of the phonological processes foundA in selected Nigerian children’s spoken English. Their utterances were examined forR these processes which were in turn isolated and put into groups, the frequency ofI tBhese processes were determined and their percentages were calculated usinLg descriptive statistics. Representative instances were analysed theoretically anNd ac oustically. These analyses revealed some phonological features about NigeriaAn children’s processes in spoken English. D IB A OF TY RS I VE UN I 188 CHAPTER FIVE SUMMARY, CONCLUSION AND RECOMMENDATIONS RY 5.0 Introduction A This chapter contains the conclusion of the research. It presents the summRary of the study and its findings, recommendations, limitations of the study, contributIioBns to knowledge and suggestions for further studies. L 5.1 Summary of the study N This study set out to examine children’s spoken EnglisAh for phonological processes and the constraints rankings responsible, with a viewA to Dtracing their language development. For this purpose, a purposively designed text Bwas read aloud by one hundred schoolchildren in Oyo and Lagos States, and two childreIn were also observed for a period of six months. Their audio-recorded utterances wFere transcribed and instances of phonological processes were isolated and subjected Oto descriptive statistics. Optimality Theory provided the theoretical tool to explaYin t he constraints rankings responsible for the emergence of the processes. LanguageT analytical computer software, PRAAT, was also used to carry out acoustic analyses oIf some of the processes. The processes were examined descriptively, rather than presScriptively so that no common process was labelled erroneous. 5.2 ESumRmary of findings FIoVllowing the research objectives as well as perceptual, theoretical and acoustic analyses, Nthese observations were made about Nigerian children’s spoken English. U 5.2.1 Phonological processes in the children’s language use The process of coalescence in the children’s utterances occurred much more in vowel segments than in consonant segments. This presents a new perspective to what exists in the literature about Nigeria’s variety of English. Also, there was a predominant substitution of dental fricatives with alveolar plosives, and dental fricatives were mainly elided when they 189 occurred contiguously with alveolar plosives. Also, dental fricatives that occurred singly were substituted regardless of where they occurred within a syllable. This finding corroborates what exists already in the literature that dental fricatives are a relatively difficult class of sounds for many Nigerians to produce spontaneously. Assimilation features occurred beyond the lexical boundaries and was predominaYntly regressive, only very few instances of progressive assimilation were recorded. This Rpresents a counter-argument to what exists in the literature that assimilation is only rAegressive in spoken English in Nigeria. Also, consonants were deleted for the purpoRse of consonant cluster reduction and juncture prosody. IB All three central vowels were predominantly strengthened to b eLcome either front or back vowels and this finding corroborates the given knowledge Nin existing literature that central vowels are usually strengthened among many NigeriAan speakers of English. Also, there appeared to be no clear modality for the childrenD’s preference for diphthongs. The front and back close vowels experienced gliding inAto semivowels in some contexts. In the longitudinal data, the youngBer participant featured intense phonological simplification like whole syllable del etIions, numerous cluster reductions, and intra-word devoicing and stopping. HoweverF, the older child displayed more adult-like and complex processes like epenthesis, gl idOing, and vowel strengthening. These complex processes were also characteristic of theY cross-sectional data. Although some proIcTesses were found throughout the data, their functions and instances varied based onS the participants’ ages. For instance, at the outset of language acquisition, deletioEn wRas carried out to remove codas, simplify consonant clusters and remove whole sylVlables, while the older children used this process to reduce strings of phonemes beyond thIe word boundary. UN5.2.2 Constraints ranking for the processes The children’s grammar favoured markedness constraints over faithfulness constraints which permitted the production of peculiar processes. However, as the children grew, their phonological processes became more similar to that of adults which also meant that the same could be said of the constraint rankings. This finding corroborates some Optimality 190 Theoretical studies on children’s phonological grammar. These markedness constraints were also responsible for the production of the children’s target phonological processes because their ambient language, Nigerian English, is, in some ways, different from the Received Pronunciation. In the children’s phonological grammar, only vowels were permitted to coalesce andY the constraints ranking responsible for this was *C VOWOMPLEX » *SCHWA » *CORMPLEX. Consonants were prohibited from coalescence and the ranking of constraints thAat helped to achieve this was UNIF » *COMPLEX » MAX. For the realisation of substRitution of dental fricatives, *FRICATIVE was often ranked very highly to achieve thisI pBrocess. Furthermore, in the achievement of monophthongisation, NO DLIPHTHONG ranked higher than MAX, while in the process of deletion, *COMPLEX wasN ranked above MAX. The central vowels were always strengthened through the procAess of vowel strengthening, thus, *SCHWA ranked higher than all the faithfulness coDnstraints like αF. For the achievement of gliding, the participants ranked *COMPLEXVOWA above αF. Epenthesis thrived by ranking *SS very hBighly, while devoicing emerged through the joint predominant ranking of FIN DEV aFnd *V I OICEDOBS]σ. 5.2.3 Language developm eOnt At the outset of languagYe acquisition, phonemes whose places of articulation are towards the labial region wIerTe acquired first. Also, voiceless phonemes were favoured over voiced ones and there Swas no distinction between pure vowels and diphthongs, rather, pure vowels were usedR generally. Furthermore, onsets in monosyllabic words were more easily articulaEted than their codas or consonants in polysyllabic words. BIyV the age of one year and nine months, enough phonemes would have been acquired for Nbasic expressive communication skills. Between the fourth and fifth year, typically U developing children would have begun to strengthen their phonemes and produce complex syllables in their utterances. At about this time also, phonological processes would have begun to take after adults’ processes. This shows that language development among typically developing Nigerian children bears a great similarity to the developmental pattern of children in developed countries. 191 There is evidence in the study to show that participants’ acquired words were amassed through interaction with, and imitation of adults who were around them. Nuances from Yoruba language and processes peculiar to it were found in aspects of the data, however, this were quite minimal. A great deal of vocabulary from surrounding adults was found in their utterances too. In other words, the ambient language used in their homes and school influenced what the children acquired. Thus, there are connections between the childrYen’s language acquisition and their linguistic environment. Also, gestures that complRemented language use were employed. These explain why phonological processes bAecome more similar to adults’ as the years pass by. This finding is not accorded eBnougRh recognition in child language studies, and is in fact relegated by renowned reseaIrchers who claim that innateness hypothesis provides the sole explanation for langua geL acquisition in children. 5.3 Conclusion AN Children’s phonological processes have been foDund to be very similar to that of adults around them. This is especially so towards tAhe end of their language acquisition phase. During the outset of acquisition howevBer, the phonological processes are usually very dissimilar from that of adults becaus eI they are still learning to adequately master and manipulate their speech organs. FThus, the products of their constraints rankings are different from that of adu ltO speakers. Children favour markedness constraints over faithfulness constraints. Y 5.4 RecommenIdTations To further the dSiscussion on phonological processes in Nigerian children’s spoken English, especially Rwith a view to tracing their language development, the following suggestions are profferEed: NI V 1. Parents’, guardians’ and teachers’ language command, education qualification and U ethnic affiliation should be brought into consideration in the data collection phase because these people have an influence on children’s language acquisition. 2. Data collection should cover children’s language use from acquisition stage till the end so that their developmental progression can be traced, analysed and theorised. Future researchers may do this on a year-by-year basis. This way, each child’s progress can be recorded and monitored to arrive at a generalisation for describing 192 child language development in Nigeria. Also, the transition between the phases of acquisition is recommended for further study. 3. While elicited data and cross-sectional collection are good, longitudinal and naturally occurring utterances should also be focused on because these are likely to contain more processes, which are also in their naturally-occurring state. Y 5.5 Limitations of the study R First, the process of data collection was strenuous and time-demanding, partlAy because it involved children. Also, this was the case because the intention was Rto create a very conducive and comfortable environment for the children so thaItB whatever data was collected would be research-worthy. Despite this however, the sLchoolchildren’s utterances were more slow-paced than expected so that phonemes wer e distinctly produced from others. Thus, there were fewer processes than expected. AlsNo, the research stage of listening to the collected data proved monotonous and arduDous.A Findings from the longitudinal data that waAs indeed available were not enough to be generalised because they involved onlyI twBo children. In addition, it was noticed that some of the children read aloud differenFtly from what was in the research instrument and this affected the data collection toO some extent. Acoustic analysis could Yneit her be carried out on all the instances of the process, nor could it be carried out on one instance for all the participants. Thus, acoustic analysis was carried out on only a repreIseTntative few. Also, the research is limited in the geographical regions it covers – onlyS Oyo and Lagos States were sampled. Finally, the research instrument was not an EofficRially standardized one, rather, it was newly designed just for this study, and may notV be devoid of possible problems. N5I.6 Contributions to knowledge U Here are the study’s specific contributions to literature: 1. Cogno-interactionist theory of language acquisition has been proposed as a more encompassing explanation for the different stages and processes that children undergo to achieve eventual language acquisition. 193 2. Combining longitudinal and cross-sectional procedures for data collection is a richer way for carrying out empirical research because it affords researchers a more well-rounded view of the data. Also, both complement each other and give research results more authenticity. This has been shown in the research and is therefore, encouraged for future studies. 3. This research has unveiled the importance of embracing the language domYain in the description of children’s development in Nigeria. ThereforeA, theR Federal Ministry of Health and the Medical and Dental Council of Nigeria are encouraged to ensure that Nigerian hospitals and their ImBedic Ral personnel are equipped with the knowledge and practice of how language domain can also be used to describe and track children’s overall develo pLment. 4. Child-peculiar phonological structures constituNte a small fraction of the reasons responsible for Nigerian children’s prefereAnce for markedness constraints over faithfulness constraints in phonologDical grammar. Rather, the distinct peculiarities of Nigerian English wAhich the children eventually come to acquire are responsible for a major itIy Bof these markedness constraints. 5.7 Suggestions for further stFudies This study is a descriptive rOesearch and has explored some phonological processes in Nigerian children’s spokYen English. It has focused on selected children in Lagos and Oyo States and is thus lIimTited in this regard. Therefore, the following areas are suggested for further studies:S 1. MoRre south-western states need to be covered to provide a more well-rounded study Eon Yoruba Nigerian children’s English phonological processes. Also, more IV researchers can explore other ethnicities in Nigeria as well as other varieties of N English that these ethnicities engender. U 2. Also, there is a need to carry out more longitudinal studies on a larger number of children so as to further trace language development. This could provide future studies with more language features for Nigerian children’s developmental milestones. 194 3. In the aspect of consonant substitution, only dental fricatives were the centre of focus, future research could include more segments which have been proved to be difficult for Nigerians. Also, all the diphthongs as well as the triphthongs could be tested among these children. 4. Future studies could also collect sociolinguistic information about participants to see how these variables combine to influence children’s phonological processesY and language development. AR R B L I N DA IB A O F IT Y S VE R UN I 195 References Ademola-Adeoye, F. 2019. A phonological description of the English language of selected 2-4 year-olds in educated families in Lagos. Lagos Review of English studies: A Journal of Language and Literary Studies 18.1:1-28. Adeniyi, K. and Adeniyi, T. 2017. Consonant sequence reduction in child phonoloYgy. Ghana Journal of Linguistics 6.1:22-48. R Akanbi, F. O. M and Anyarsor, C. 2014. Growth monitoring: the key to chAild survival strategy in Nigeria. Journal of Research in Nursing and MidwiferyR 3.2:31-38. Akpan, E. E. 2006. Assimilatory processes in Ibibio child phonologyI. SBelected Proceedings of the 36th Annual Conference on African Linguistics, OL. F. Arasanyin and M. A. Pemberton. Eds. Somerville: Cascadilla ProceedingNs Project. 27-36. Alerechi, R. I. C. 2011. Consonant substitution inA child language (Ikwere). African Linguistics on the Prairie: Selected PapeDrs from the 45th Annual Conference on African Linguistics, J. Kandybowicz, TA. Major, H. Torrence and P.T. Duncan. Eds. Berlin: Language Science Press. 1B35-155. Ambrose, M. J. 2010. Overextens IFion in language development: a case study of a Gwong child. Master’s Project. Dept. of English and Literary Studies. Ahmadu Bello University. 1-135. O Augustyn, A., Bauer, YP., Duignan, B., Eldridge, A. Gregersen, E., Luebering, J.E., McKenna, IAT., Petruzzello, M., Rafferty, J.P., Ray, M., Rogers, K., Tikkanen, A., WallenfSeldt, J., Zeidan A., and Zelazko, A. 2015. Britannica Encyclopaedia. ChiRcago: Encyclopaedia Britannica. 401-404. Baker,E L. 1979. Comprehension monitoring: identifying and coping with text confusions. IV Journal of Reading Behavior. 11.4: 365-374. NBall, M. J. 2016. Principles of clinical phonology: theoretical approaches. New York: U Routledge. Bowen, C. 2011. Table 2: Phonological processes. Retrieved on Wednesday, June, 2019 from http://www.speech-language-therapy.com/ Bowen, C. 2011. Table 3: Elimination of phonological processes. Retrieved on Wednesday, June 12 from http://www.speech-language-therapy.com/ 196 Brightstone, R. E. 2000. Language acquisition: perspectives, development and relation to thought. Jones and Bartlett Publishers LLC 71-93. Brown, R. 1973. A first language: the early stages. Cambridge: Harvard University Press. Campbell, R. and Wales, R. 1970. The study of language acquisition. New Horizons in Linguistics. J. Lyons. Ed. New York: Penguin Books. 1-13. Y Chapitré, P. P. 2005. Phonological processes. The practicality of speaking. RMexico: Duncan-high Publishers. 89-103. A Chomsky, N. 1959. Verbal behaviour by BF Skinner: a review. LanIguBage R 35:1: 26-58. Chomsky, N. 1996. Language and evolution-reply. New York R eLview of Books 43.2: 41-41. Coleman, L. M. 2013. Language and the evolution of Nidentity and self-concept. The Development of Language and Language ReseAarchers. Psychology Press 331-350. Côté, M. 2000. Consonant cluster phonotactics: a Dperceptual approach. Ph.D. Thesis. Dept. of Linguistics and Philosophy. MaBssacAhusetts Institute of Technology.1-96. Crain, S. and Lillo-Martin, D. 19F99. A In introduction to linguistic theory and language acquisition. Massachusetts: Blackwell Publishers. Cruttenden, A. 2001. Manc uOnian intonation and intonational representation. Phonetica. 58.1-2: 53-80. Y Davis, S. 1995. EmIpThasis spread in Arabic and grounded phonology. Linguistic Inquiry. 465-498S. DopemEu, ER. A. 2015. Early childhood English language acquisition of Hausa children IV living in rural communities of Kano state Nigeria. The European Conference on N Education official conference proceedings. The International Academic Forum 1-U 13. Duncan, T. S. 2013. Language acquisition (in early childhood). PowerPoint presentation. Dept. of Linguistics. University of Alberta. Eren, O. 2015. Children’s acquisition of the phoneme /r/ in Gaziantep province. Journal of Child Language Acquisition and Development 3.4: 177-183. 197 Etim, V. E., Dada, O. A. and Bassey, O. 2018. A prevalence of phonological disorder in children with specific language impairment. International Journal of Cognitive and Language Sciences 12.6: 267-270. Eze, J. N. Oguonu, T., Ojinnaka, N. C. and Ibe, B.C. 2017. Physical growth and nutritional status assessment of school children in Enugu, Nigeria. Nigerian Journal of Clinical Practice 20.1: 64-70. RY Fernald, A. and Marchman, V. A. 2006. Language learning in infancy. HAandbook of Psycholinguistics. M.J. Traxler and M.A. Gernsbacher Eds. 2ndR ed. San Diego: Elsevier Incorporated. 1027-1071. IB Fernández, E. M. and Cairns, H. S. 2011. Fundamentals of ps ycLholinguistics. New York: John Wiley and Sons. Field, J. 2004. Psycholinguistics: the key concepts. New YNork: Routledge. Field, J. 2006. Psycholinguistics: a resource booDk foAr students. Oxon: Routledge Taylor and Francis Group. A Garnham, A., Garrod, S. and Sanford,I AB. 2006. Observations on the past and future of psycholinguistics. Handbook of Psycholinguistics. M.J. Traxler and M.A. Gernsbacher Eds. 2ndO ed. SFan Diego: Elsevier Incorporated. Graven, S. N. and BrownYe, J . 2008. Auditory development in the fetus and infant. Newborn and Infant NuTrsing Reviews. 8.4: 187-193. GuolaugsdottirS, E.I B. 2016. Children’s first language acquisition: what is needed for chiRldren to acquire language? Bachelor of Arts Long Essay. Dept. of English. ESchool of Humanities, Haskoli Islands. 1-85. HIaVll, N. 2011. Vowel epenthesis. The Blackwell companion to phonology: phonological N processes. Eds. M. van Oostendorp, C. J. Ewen, E. Hume and K. Rice. Oxford: U Wiley-Blackwell. 3: 1576-1596. Hendriks, P. and Spenader, J. 2006. When production precedes comprehension: an optimization approach to the acquisition of pronouns. Language Acquisition. 13.4: 319-348. Hepper, P. G. and Shahidullah, B. S. 1994. Development of fetal hearing. Archives of Disease in Childhood. 71: F81-F87. 198 Hutauruku, B.S. 2015. Children first language acquisition at age 1-3 years old in Balata. Journal of Humanities and Social Science. 20.8: 51-57. Hyams, N. and Orfitelli, R. 2015. The acquisition of syntax. Handbook of Psycholinguistics. Eds. H. Cairns and E. Fernández. New York: Wiley/Blackwell Publishers. Ibrahim, S. S. and Ibrahim, H.I. 2016. Child language acquisition in relationY to reduplication: a case study of 6 (six) two-year-old Hausa children. InterRnational Journal of Humanities and Social Sciences. 5.6: 113-124. A Josiah, U. E. and Soneye, T. 2014. Peculiar patterns of juxtapositioBnalR assimilations in educated Nigerian spoken English variety. Linguistik OnLlineI 15.1: 125-142. Jowitt, D. 1991. Nigerian English usage: an introduction. Lag os: Longman. Kager, R. 2001. Optimality theory. Cambridge: CambridgeN University Press. Kerswill, P. 1985. A sociophonetic study of connecAted speech processes in Cambridge English: An outline and some data. DCambridge Papers in Phonetics and Experimental Linguistics, 4:45-69. A Kramiņa, I. 2000. Lingo - Didactic theorIieBs underlying multi-purpose language acquisition. University of Latvia. 87-98F. Kroon, S. 2003. Mother tongOue and mother tongue education. Language Education. J. Bourne and E. Reid. Eds. 35-48 Y Laver, J. 1968. AssimTilation in educated Nigerian English. ELT Journal. 22. 2: 156-160. Levelt, W. J. MS. 2I013. A history of psycholinguistics: the pre-Chomskyan era. Oxford: OxRford University Press. Locke,E J. L. 2009. Evolutionary developmental linguistics: naturalization of the faculty of IV language. Language Sciences. 31.1: 33-59. NLust, B. C. 2006. Child language: acquisition and growth. Cambridge: Cambridge U University Press. Matthews, A. 1996. Linguistic development. International Journal of Child Development. 7.6: 131-142. 199 McCarthy, J. J. 1997. Process-specific constraints in Optimality Theory. Linguistics Department Faculty Publication Series. University of Massachusetts Amherst 1- 23. McCarthy, J. J. 2007. What is optimality theory? Linguistics Department Faculty Publication Series. Massachusetts: University of Massachusetts Amherst 1-28. Menn, L. and Vihman, M. 2011. Features in child phonology: inherent, emergentY, or artefacts of analysis? Where do phonological features come from? G.N. CRlements and R. Rudouane Eds. London: John Benjamins Publishing Company.A 261-301. Nelson, K. 1977. First steps in language acquisition. Journal of the ABmerRican Academy of Child and Adolescent Psychiatry. 16.4: 563-583. I Nolan, F. J. and Kerswill, P. E. 1990. The description of c onLnected speech processes. Studies in the pronunciation of English: A commemNorative volume in honour of A.C. Gimson. S. Ramsaran Ed. London: Routledge A295-316. O’Grady, W. and Cho, M. 2001. First languagDe acquisition. The Study of Language Acquisition. 410-431. A Oladipupo, R. O. 2014. Aspects of conIneBcted speech processes in Nigerian English. Sage Open 4.4: 1-6. Olarewaju, G. B. 2018. CoOnstraFint interaction in the acquisition of English syllable structure by Yoruba- English bilingual children. Master’s Project. Dept. of English. University ofT IbaYdan. Olarewaju, G. BS. anId Sunday, A. B. 2020. Sonority and acquisition of onset clusters among YoRruba-English bilingual children. Journal of the English Scholars’ Association of Nigeria. 22.2: 26-52. OInVwub Eiko, C. D. 2011. Assessing language development in bilingual school children N between 5-8 years: a case study of Igbo/Yoruba usage in Ibadan metropolis. U Master’s Project. Dept. of Linguistics and African Languages. University of Ibadan. Piaget, J. 1970. History, theory and methods. Handbook of Child Psychology. P.H. Mussen Ed. 1: 294-346. Pike, K. 1948. Tone Languages. Michigan: University of Michigan Press. 95-165 200 Popoola, O. A. and Adeoti, A. 2016. Child welfare deprivation in rural Nigeria: a counting approach. Child Development Research. 148 – 157. Prince, A. and Smolensky, P. 1993. Optimality theory: constraint interaction in generative grammar. Cambridge, Massachusetts: MIT Press. 1-9. Quelroga, B. A. M., Rosal, A. G. C., Solva A. C. F. and Cordeiro A. A. A. 2015. AnalYysis of phonological processes in children from the metropolitan region of Recife. CEFAC. 17.5: 1449-1456. A Rahimpour, M. 2004. Developmental stages of child language. JournRal of Faculty of Letters and Humanities. 47. 190: 57-70. IB Rice, M. L. 1989. Children’s language acquisition. AmericaLn Psychologist. Kansas: University of Kansas. 149 – 156. Roach, P. and Widdowson, H. G. 2001. Phonetics. OxfAord:N Oup Sánchez-Reyes Peñamaría. Rossington, C. E. 1981. Environmental aspects ofD child growth and nutrition: a case study from Ibadan, Nigeria. Ibadan: Kluwer Academic Publishers. 5.4: 347-356. Sachs, J. Bard, B. and Johnson, M.L. 1I98B1. Language learning with restricted input: case studies of two hearing chilFdren of deaf parents. Applied Psycholinguistics. 2.1: 33- 54. O Saffran, J. R., Senghas,Y A. and Trueswell, J. C. 2001. The acquisition of language by children. PNIATS. 98.23: 12874 – 12875. Salim, J. A. andS Mehawesh, M. 2014. Stages in language acquisition: a case study. English LanRguage and Literature Studies. 4.4: 768-925. ShaeffeEr, S.F. 2020. Mother tongue and Early Childhood care and education: synergies and IV challenges. Bangkok: United Nations Educational, Scientific and Cultural N Organization. 1-19. Retrieved on 17th September 2021 from U http://www.unesco.org/open-access/terms-use-ccbysa-en Shooshtaryzadeh, F. 2015. Optimality theory and assessment of developing and disordered phonologies. Journal of Indian Speech Language and Hearing Association. Wolters Kluwer: Medknow. 29.2: 13-20. Skinner, B.F. 1957. The experimental analysis of behaviour. American Scientist. 45.4: 343- 371. 201 Stork, F. C. and Widdowson, D. A. 1974. Learning about linguistics. London: Hutchinson Ltd. Sunday, A. B and Olarewaju, G. B. 2020. Acquisition of complex coda and sonority among selected bilingual Nigerian children. Journal of Child Language Acquisition and Development. 8.4: 144-161. Sureshkumar, N. 2002. Evolving a new English language acquisition programmeY at primary level - a cognitive interactionist approach. Ph.D. Thesis. DeApartRment of English, University of Calicut. 41-101. Sutomo, J. 2012. English phonological processes: a study of generativBe phRonology theory. Jurnal Dinamika Bahasa dan Budaya 7.2: 56-64. I Szczegielniak, A. 1982. Language Acquisition. PowerPoint PLresentation. Retrieved on June 12, 2019 from https:www.scribd.com/dNocument/347492185/Language- Acquisition-ppt A Tanenhaus, M. K. 1989. Psycholinguistics: anD overview. Linguistics: the Cambridge Survey. 3: 1-37. A Utulu, D. C. 2014. MonophthongisatiIonB and vowel lengthening processes in educated Urhobo English: a moraic acco unt. Arts and Design Studies. 26:45-54. Vygotsky, L. 1978. InteractOion bFetween learning and development. Readings on the development of child ren. 23.3: 34-41. Y Weisser, M. 2005. ICToalescence. English Phonetics and Phonology. Retrieved on 28th August, 2021 from maRrtinwSeisser.org/courses/phonetics/connect/coalescence.html. WeVlls, EB. and Stackhouse, J. 2016. Children’s intonation: a framework for practice and I research. West Sussex: John Wiley and Sons Ltd. UNWundt, W. 1900. Die sprache. 2nd ed. Leipzig: Kröner. Yousofi, N. and Ashtarian, S. 2015. Protoword and true word production in children of 9- 36 months: the case of a Kurdish speaking child. Journal of Child Language Acquisition and Development. 3.4: 227-255. 202 Zaščerinska, J. 2010. Language acquisition and language learning: developing the system of external and internal perspectives. 52nd International Scientific Conference of Daugavpils University, Daugavpils, Latvia. 1-8. Zsiga, E. C. 2003. Articulatory timing in a second language: evidence from Russian and English. Studies in Second Language Acquisition. 25.3: 399-432. Y Zsiga, E. C. 2011. Local assimilation. The Blackwell Companion to PhoRnology: Phonological Processes. M. van Oostendorp, C. J. Ewen, E. Hume aAnd K. Rice. Eds. Oxford: Wiley-Blackwell. 3: 1919-1944. R B L I AN BA D I O F Y SI T VE R I UN 203 APPENDIX 1 Constraints of phonological processes The phonological process-related constraints that were used in the course of analysis in this study are presented below. Y 1. *C]σ - Syllables must not have codas. (Borowsky, 1989: p. 145-166) R 2. *COMPLEX: a syllable has no more than one consonant in a margin (WhAeeler, 2007: p. 3) R 3. *C VOWOMPLEX : No string of vowels within a syllable (GreeIn,B 2010: p. 101) 4. *DIFF[SON]: No sonority differences between onset consoLnants (Pons, 2005: p. 13) 5. *F-ONS: A fricative cannot occupy the onset positiNon (P ater and Barlow, 2003: p. 487-526) 6. *FRICATIVE: Segments may not be [+contD, –soAn] (Pater and Barlow, 2003: p. 487-526) 7. *SCHWA: The realisation of a central Avowel is banned (Wheeler, 2007: p. 8) 8. *VOICEDOBS]σ – this may be r esIpBonsible for coda obstruent devoicing (Mustafawi, 2006: p. 72) F 9. AGREE(place): an inp uOt and output must have identical features of place (Lamont, 2015: p. 54) 10. A YGREE(VOICIET): Assign one violation mark for every pair of adjacent consonants whose mSembers differ in their specification for voice (Lamont, 2015: p. 54). 11. FINRAL VOICING: Phonemes in syllable coda must be voiced (Lamont, 2015: p. 55) 12. EIDENT(MANNER): Assign one violation mark for every manner feature in the input that IV is different from its corresponding manner feature in the output (Lamont, 2015: p. N 54). U 13. IDENT(PLACE): Assign one violation mark for every place feature in the input that is different from its corresponding place feature in the output (Lamont, 2015: p. 54). 14. IDENT-IO(VOICE): Assign one violation mark for every voice feature in the input that is different from its corresponding feature in the output – underlying specifications for voice should be respected (Marc van Oostendorp, 2011: p. 3) 204 15. MAX: No deletion. Every phoneme in the base should have a correspondent in the output (Wheeler, 2007: p. 3) 16. NODIPHTHONG: Diphthongs are prohibited (Green, 2010: p. 101) 17. UNIFORMITY: It penalises segmental coalescence, no coalescence. No element of S2 has multiple correspondents in S1 (Wheeler, 2005: p. 62) 18. αF: Correspondent segments in input and output have identical values for [voiYce], [high] and [cont] (Mustafawi, 2006: p. 76) R 19. σCONT: The onset of a syllable must not be of greater sonority than the Alast segment (Pons, 2005: p. 13) BR L I N AD A IB OF Y SI T VE R I UN 205 APPENDIX 2 Reading Miss Jane had a home with a gate. “I will plant a vine next to the gate,” said Miss Jane. It will add a bit of blue. I like blue.” Miss Jane went to the vine Y shop. She got a vine. It was in a pot. R Miss Jane set the vine in the sun. She fed the vine. A She gave it a lot of water. Then, there was a bud. R “It is not a vine”, said Miss Jane. It is a rose. AL reId B, red rose. In time, the rose was big. “I am glad I got Na rose,” said Miss Jane with a smile. Miss Jane cut the stem. SAhe set the rose in a vase. The rose lit up Miss Jane’s home BA . D *I * * At my second year birthFday, grandma’s yummy cake was all I ate. Later, dOaddy tuned in to Tom and Jerry. Then my friends sa ng the birthday song while mum combed my haIirT. T Yhe last yo-yo was given to my sister. * * * GrandpaR anSd grandma will buy new textbooks and biscuits for us this wEeekend because school resumes in July. Sophia and I have aIlsVo marked our calendar because we will celebrate our birthday Nthis month. We will use our calculator to calculate the remaining U days till the celebration. All our favourite cartoon characters on television will be present. We will walk around under an umbrella and even do some magic on our friends. 206 Reciting 1. Rain, rain go away Come again another day Y Little children want to play R Rain, rain go away A 2. Aeroplane, father aeroplane/2x R I want to be a pilot, Aeroplane IB Like this, like this L Like this, like this, like this DA N A F IB O IT Y RS IV E UN 207