Journal of Linguistics and Language in Education Vol 8, Number 1 (2014): 22-36 Phonological Adaptation of Loanwords into Mad Theory Account a: An Optimality Greg Obiamalu* and Linda Nkamigbo** Abstract This paper, using Optimality Theory (OT) as a framework, investigates the phonology of loanwords in Ma da , a Benue-Congo language spoken in Nasarawa state, North Central Nigeria (Williamson & Blench, 2000). Data for the study were obtained from two sources: Mada- English dictionary and audio recordings of relevant corpus from three native speakers of Ma da . The standard OT view of loanword phonology assumes that it is the host language grammar that acts on the foreign words by selecting the right output form from a vast majority of candidates. The paper examines the strategies that Ma da adopts in the process of adapting words borrowed from other languages into it as well as the constraints hierarchy preferences. The results show that the phonological changes which foreign words undergo when borrowed into Ma da are evident in a range of phonological processes such as vowel deletion, coda simplification, cluster simplification, structure preservation, and syllable deletion. In all, it is observed that, to a large extent, the loanwords violate the syllable structure and Mada phonotactics; but in order to preserve its structure, Mada uses constraints that require the output material to be independent of input. Keywords: Mada , loanwords, optimality Theory, constraints Introduction Haspelmath (2009) defines a loanword as a word that at some point in the history of a language entered its lexicon as a result of adoption from another language that may or may not be related to the receiving language. This phenomenon is traditionally known as borrowing. But since borrowing has other semantic connotations like "taking and returning something?, other terms such as "transfer? or "copying? are used to refer to the same phenomenon. In loanword phonology, loans are required to conform to the host language grammar and still remain faithful to the donor language. This is especially difficult when the structure of the host language and the donor language are very different. Loans are meant to agree with the segmental and phonotactic constraints of the host language. The interplay between the conformity to host grammar and faithfulness to the donor language results in conflicting forces. In an attempt to resolve this conflict, Kenstowicz (2005:1) observes that "a model of input-output mapping that formalizes the resolution of conflicting forces, driving the input towards specific output targets *Lecturer, Department of Linguistics, Nnamdi Azikiwe University, PMB 5025 Awka, Nigeria, E-mail: gregobiamalu@yahoo.com **Lecturer, Department of Linguistics, Nnamdi Azikiwe University, PMB 5025 Awka, Nigeria, E-mail: linkamigbo@gmail.com seems a priori better suited to model this aspect of linguistic competence than the ordered-rule framework of traditional generative phonology". This paper, therefore, uses Optimality Theory (OT) to investigate the phonology of loanwords in Ma da , a Benue-Congo language spoken in Nasarawa state, North Central Nigeria (Williamson & Blench, 2000). Ma da borrowed extensively from Hausa and English, with a few loans from Shugbu and Yoruba. Ma da loanwords have been documented in a Ma da -English dictionary, thus providing an easily accessible wealth of material for this analysis. Additional data on loanwords were elicited from three native speakers of Ma da . The rest of the paper is organized as follows: Section Two presents the review of related literature on loanword phonology with particular attention paid to the works that use Optimaliy Theory in their analysis of loanwords. For a better understanding of how loanwords are adapted into Ma da , Section Three sifts through Ma da phonology. In Section Four, the data for this study, Ma da borrowings, are presented. Section Five focuses on the analysis of the data. Section Six gives a summary of the findings, and some unresolved issues that could lead to further research. Some Previous Studies on Loanword Phonology There is no existing literature on Ma da loanword phonology, to our knowledge. However, a large body of literature on loanword phonology in other languages exists. Haspelmath and Tadmor (2009) is a book on loanwords in different language families of the world. The closest languages discussed in the book are Hausa (Chadic) and Kanuri (Saharan) by Ari Awagana, H. Ekkehard Wolff and Doris Löhr. Burenhult?s (2001) research sifts through the phonological treatment of loanwords in Jahai. The phonological changes in Jahai loanwords are organized into four categories: phonetic adaptation such as preplosion of word-final nasals, nasalisation of word-initials, and palatalisation of /s/; phonemic replacement such as substitution of a final syllable /a/, and a word-final glottal stop; reorganisation of syllable structure such as closure of syllables, and reinterpretation of word-medial vowel sequences; and relocation of stress. Golston and Yang (2001) study the loanword phonology of White Hmong using OT. The research reveals the following: exact borrowings, which they attribute to the faithfulness constraints MAX and DEP (McCarthy & Prince, 1995); structure preservation such as the replacement of the French [?] and the English [?] by Hmong [l]; satisfaction of the constraint "ONSET? by the insertion of the glottal stop [?]; and satisfaction of NO CODA by consonant deletion. Yip (2002) looks into the nature of Cantonese loanword phonology from OT viewpoint. The paper has two sections. The first section reviews consonant adaptation, in which excess consonants are either salvaged by epenthesis or deleted. Yip adds MIMIC, a faithfulness constraint specific to loanword adaptation, to the usual OT view of loanword adaptation. MIMIC is the OT instantiation of active loanword incorporation, and enforces faithfulness to the percept (Yip, 2002:2). The second section of the paper looks at vowel adaptation, and explores the extent to which acoustic similarity determines the choice of vowels. Some of the findings include: [a:] or [e:] are found in the obligatorily long open output syllables, and before nearly all nasals. Before stops, the vowel is always short and the reflex is usually [I], with one [?] before the only [t]. There is also one [I] before [m]. Aziza and Utulu (2003), commenting on the phonological nature of Urhobo and Yoruba (Benue-Congo languages) loanwords from English, identify vowel lengthening as a strategy through which Yoruba adapts English words. Oyebade (2006), using OT, agrees with Aziza and Utulu, but reveals that the motivation for Yoruba?s vowel lengthening in words of English provenance is Yoruba?s desire to preserve the prosodic structure of the syllable(s) of such word(s) as they come from their source. Akinlabi?s (2008) proposal on Yoruba loanword phonology has the following points of interest: The primary prosodic changes are onset simplification, coda simplification, and the syllabification of diphthongal glides as onsets. The strategies for these simplifications include cluster reduction, vowel epenthesis, and [h] epenthesis. In addition, codas are also deleted and diphthongs are "assimilated". Phonologically, the preference is for retaining the original place of articulation in consonants and changing the manner of articulation to conform to the Yoruba inventory. Stress is approximated to a falling tone, with the high tone on the stressed vowel and the following vowels bearing low tones. If stress does not occur on the initial syllable in the source word, all syllables before the stressed syllable may also be assigned low tones. From this review it is very clear that OT can account for loanwords. It is, therefore, more relevant to the present study than rule-based phonology is. Ma da Phonology: An Overview Consonants Ma da uses only pulmonic consonants. Unlike many Central Nigerian languages, it does not have implosives or ejectives. There are twenty-six consonants in Ma da . They include: ten plosives, three nasals, six fricatives, three affricates, one trill, and three approximants. Their places of articulation are: bilabial, labiodental, alveolar, postalveolar, palatal, velar, labiovelar, and glottal. Ma da consonant phonemes are presented in Table 1 below. Table 1: Ma da Consonant Chart Bilabial Labiodental Alveolar Post alveolar Palatal Velar Labiovelar Glottal Plosive p b t d c ? k g kp gb Nasal m n ? Fricative f v s z ? h Affricate ts t? ? Trill r Lateral l Approximant j w Vowels u Close i Close-mid e o Open-mid e ? ? Open a Figure 1: Mad a Oral Vowel Chart Ma da vowels are distinguished by height, fronting as well as nasality. There are thirteen vowels in Ma da : eight oral /i e ? u o ? ? a/, and five nasalized /i ? u ? a / vowels. Front Central Back Close Central Back u æ Close-mid Open-mid e Open a Figure 2: Ma da Nasalized Vowel Chart Tone Ma da is a register tone language because it uses tone contrastively on lexical items. It is a three-tone language with the syllable as the tone-bearing unit. Price (1989) identifies two irregular contour tones in the language, in addition to the three basic ones. Ma da tones are represented below using the popular convention in the language: High = Example /dó/ Gloss "market? Mid Low = = unmarked /t?i/ /bwa / "to move? "to pour out? Rising Falling = = ? ˆ /fe/ /?? / "to maltreat someone? "instead? Instances of tonal permutations in Ma da words are as shown below: Example Gloss (1) a. /?a / "to rebuke a younger person? b. /?a / "not completely ripe? (2) a. /k?de / "like? b. /k?de / "something remaining (3) a. /mf?/ "hunger? b. /mf?/ "foam? (4) a. /r?/ "thirst? b. /re/ "front of the house? Ma da Syllable Structure The Ma da syllable structure is schematized below. (ONSET) PEAK (CODA) Figure 3: Ma da Syllable Structure The use of dotted lines shows that the onset (ONS) and the coda are optional, while the thick line depicts the obligatory nature of the peak in every Ma da syllable. The peak is usually a vowel but it may be a syllabic nasal in some restricted environments. Basic syllable types in Ma da are: V, CV, and CN. The V Syllable Type This syllable type is found only in word initial position in some bi/multisyllabic words. It occurs in a few monosyllabic functional words such as in prepositions, and conjunctions, as the example below indicates: Example Gloss (5) a. [ã] an "and? b. c. [?] [? ] ? en "ate? "yes? (in response to a question) The CV Syllable Type This is the most common syllable type. Some examples are listed in (6) below: Example Gloss (6) a. /ts?/ "to vomit? b. /?i / "leprosy? c. /ga/ "shoulder? d. /se/ "transplant? e. /to/ "to crow? The CN Syllable Type The onset of the CN syllable is one of the labio-velar plosives. The nucleus is a syllabic nasal, represented as /m/, as in in (7) below. Example Gloss (7) a. /kpm/ "kapok tree? b. /gbm/ "canoe? Borrowings in Mad a: The Data The largest body of data is from Hausa and English loans into Ma da , though a few loans from Shugbu and Yoruba are also analysed in this paper. These loans come from two sources: the Ma da -English dictionary and recordings from three native speakers of Ma da . The loanwords are organized into six categories. Exact Borrowings Here, the sounds of the loanwords are the sounds that Ma da and the donor language(s) share. Examples are presented in Table 2 below. Table 2: Examples of Exact Borrowings Ma da form Source form Source language English gloss bel ?be la bel ?bela Hausa cattle egret bi bi Shugbu cobra cinci cinci Hausa fried sweet dough cund??ma cund??ma Hausa chief of people kal? ba kal? ba Hausa bottle ku ka ku ka Hausa baobab kwakwa kwakwa Hausa coconut wando wando Hausa trousers ta ki ta ki Hausa fertilizer lamu ka lamu ka Hausa typhoid All the sounds in the loanwords above are found in Ma da . Initial Vowel Deletion Initial vowel deletion is observed in some loanwords. The initial vowels of some source words are deleted when they are adapted into Ma da . Instances of vowel deletion are given in Table 3. Table 3: Examples of Initial Vowel Deletion Ma da form Source form Source language English gloss da ada Hausa cutlass kara akara Yoruba bean cake kwat i akwa ti Hausa box l? ba sa al ? basa Hausa onion le de alede Hausa pig gogo agogo Hausa wrist watch ma nuelu Im?'nuel English Emmanuel le s?ba to IlIz?'biO English Elizabeth Final Vowel Insertion or Consonant Deletion Ma da simplifies coda either by vowel insertion or consonant deletion. The vowel insertion or consonant deletion could be seen in the words in Table 4. Table 4: Examples of Final Vowel Insertion or Consonant Deletion Ma da form Source form Source language English gloss ma nuelu Im?'nuel English Emmanuel le s?ba to IlIz?'biO English Elizabeth site?onu 'steI?n English station b?su b?s English bus kokusu 'k?k?s English caucus bent?i bent? English bench nazareti naz?reO English Nazareth sida 'sid?r English cedar poli p?'lis English police kaseti k?'set English cassette fonu f??n English phone ofisi '?fIs English office otomati ?t?'mætIk English automatic Cluster Break-up In Ma da , consonant clusters are salvaged by vowel insertion. Vowels are inserted in order to break up consonant clusters in loans. The clusters could be either at the onset or coda of a syllable. The examples in Table 5 illustrate this phenomenon. Table 5: Examples of Cluster Break-up Ma da form Source form Source language English gloss kul?bu kl?b English club kilasi kl?s English class buro?i br?? English brush sitovu st??v English stove digiri dI'gri English degree tebulu 'teIbl English table Structure Preservation In an attempt to preserve Ma da phonology, loanword sounds that are lacking in Ma da phoneme inventory are replaced with the ones that exist in the language. For instance, the following sounds do not exist in Ma da : /au æ ?? I aI eI O ? ?/. /au/ is represented by /?/; /??/ is represented by either /o/, /?/, or /a/ as the case may be; /aI/, and /eI/ are represented by /e/; /I/ is represented by /i/; /O/ is represented by /t/; /?/ is represented by either /?/ or /o/; and /?/ is represented by /d?/. Let us consider the examples in Table 6. Table 6: Examples of Borrowings Involving Structure Preservation Ma da form Source form Source language English gloss d?ro dauro Hausa millet m?ngoro mæng?? English mango mato m??t? English motor mi??n 'mI??n English mission se saI Hausa until les ?ba to IlIz?'biO English Elizabeth Ma da form Source form Source language English gloss kul?bu kl?b English club d?uelu ?oel English Joel kilasi kl?s English class d?osefu ?ozef English Joseph buro?i br?? English brush d?akbu ?æk?b English Jacob sitovu st??v English stove digiri dI'gri English degree d?la 'd?l? English dollar tebulu 'teIbl English table Syllable Deletion It is observed that Ma da deletes some syllables in certain loanwords, while trying to adapt the words. This deletion could be at the Syllable Initial Position, Syllable Medial Position or Syllable Final Position. It is not clear what motivates this deletion, but it seems to follow a general economy tendency in languages. It could be likened to vowel and consonant deletions which are common features of natural languages. Syllable deletion in Ma da loanwords is evident in the words in Table 7 below. Table 7: Examples of Borrowings Involving Syllable Deletion Ma da form Source form Source language English gloss da dama Hausa lead ba taba Hausa to touch be rube Hausa rotten ba zuba Hausa to pour out d?i d?i ra Hausa to wait ladi lahadi Hausa Sunday riri siriri Hausa slender person zinari zinari ja Hausa silver ka ra nta makaran ta Hausa school Constraint-based Account of Ma da Loanwords: The Analysis Optimality Theory (OT), first introduced by Prince and Smolensky (1993) and developed by McCarthy and Prince (1994), is a constraint-based phonological system that allows violable constraints in deriving output surface forms from underlying forms. OT assumes that linguistic items are restricted by a set of universal, mutually inconsistent and violable constraints from which an optimal surface output will be selected. Oyebade (1998), quoting McCarthy and Prince (1993), summarizes the basic assumptions and principles of the theory this way: Optimality Theory assumes that the role of a grammar is to select the right output form from among a very wide range of candidates, including at least all of the outputs that would be possible in any language whatsoever.... Language-particular rules or procedures for creating representations have no role at all in the theory and the...burden of accounting for the specific patterns of individual languages falls on the well-formedness constraints (McCarthy & Prince, 1993:4). McCarthy and Prince (1994:336) present five basic principles of Optimality Theory; the three most important ones are: a. Universality: UG provides a set CON of constraints that are universally present in all grammars. b. Violability: Constraints are violable; but violation is minimal. c. Ranking: The constraints of CON are ranked on a language particular basis; the notion of minimal violation is defined in terms of this ranking. A grammar is a ranking of the constraint set. McCarthy (2008) further stipulates how OT constraints should be formulated. According to him, OT provides the tools for investigating constraints which could be voilated. It simply states that the constraints are of two types: markedness and faithfulness. The ranking of these constraints is language specific and ranking determines the variations seen in languages. We adopt a constraint-based approach, Optimality Theory, because the constraints that comprise the native phonology can be used to understand how loanwords are nativized. This paper assumes the following constraints, which are relevant to the nativization processes in Ma da . The constraints are arranged according their ranking in Ma da loanword phonology, starting from the highest ranked to the lowest ranked. The rationale for the ranking of each constraint is explained. 1. STRPRES (structure preservation) (Kiparsky, 1982). Only contrastive sounds of the language are allowed. This is the highest ranked constraint because only Ma da phonemes are allowed in the loanwords. 2. Complex onset (McCarthy, 2002). Consonant clusters are prohibited in the onset of a syllable. This constraint is highly ranked in Ma da because the language never allows consonant clusters. This constraint paves the way for vowel insertion. 3. Complex coda (McCarthy, 2002). Consonant clusters are prohibited in the coda of a syllable. This constraint is highly ranked as well because Ma da does not allow consonant clusters. This constraint also paves the way for vowel insertion. 4. No coda (Prince & Smolensky, 1993). Syllables end with vowels. This constraint is highly ranked in Ma da because closed syllables are not found in it. Even though there are a few cases of CN syllabes as shown in the discussion of "Ma da syllabe structure? above, the N is syllabic, and therefore behaves like vowels. 5. Onset (Prince & Smolensky, 1993). Syllables begin with consonants. This constraint is high on the ranking because the CV syllabe structure is the most common in Ma da . However, there are some V syllables in the language, which makes the ranking of ONSET not too high. 6. IDENT: IO (McCarthy and Prince, 1995). The output material must be identical to that of the input. This constraint has an intermediate ranking in Ma da because even though there are loanwords, in which the input (source form) is the same as the output (Ma da form), there are many more that are different. 7. MAX-SYL (McCarthy & Prince, 1995). No syllable deletion. This constraint is also at the intermediate ranking because syllable deletion may or may not occur in Ma da loanword phonology. 8. MAX-C,-V (McCarthy & Prince, 1995). No consonant or vowel deletion. This is similar to IDENT IO. This constraint is low on the ranking because Ma da allows consonant and vowel deletion. 9. DEP-C,-V (McCarthy & Prince, 1995). No consonant or vowel insertion. This constraint is the lowest on the ranking because Ma da always allows vowel insertions to break up consonant clusters. With these constraints, the adaptation of loanwords in Ma da can be well accounted for. Let us now examine the constraints involved in the different types of borrowings using one example from each group. The relevant constraints are determined by the possible competing candidates. These constraints are ranked in line with the ranking of the constraints in Ma da loanword phonology stated above. Exact Borrowings e.g. ta ki from Hausa ta ki "fertiliser? Table 8: Exact Borrowings NO CODA > >ONS >> IDENT-IO >> DEP-SYL >> DEP-SYL NO CODA ONS IDENT-IO DEP-C, -V at.a.ki *! * * * ? ta.ki ta.kit *! * * a. ta.ki *! * * ta.ki voilates no constraint and therefore emerges as the winning candidate. Initial Vowel Deletion e.g. gogo from Hausa agogo "wrist watch? Table 9: Initial Vowel Deletion NO CODA>> ONS >> IDENT-IO>> DEP-C >> MAX-V NO CODA ONS IDENT-IO DEP-C MAX-V a. go. go *! ga. go. go * *! ? go. go * * a. go. gog *! * * * Although go.go violates the faithfulness constraints - IDENT-IO, and MAX-V, it still emerges as the optimal candidate because the contending candidate ga.go.go violates a higher ranked constraint DEP-C, which go.go does not violate. The deletion of /a/ is attributed to the presence of a constraint requiring syllables to begin with consonants. a.go.go violates a higher ranked constraint - ONS (ONSET). Vowel Insertion e.g. site?onu from English stei?n "station? Table 10: Vowel Insertion COMP ONS >> COMP CODA >> NO CODA >> IDENT-IO >> DEP-V COMP ONS COMP CODA NO CODA IDENT- IO DEP-V steI.?n *! * * ste.?on *! * * * ? si.te.?o.nu * *** ste.?o.nu *! * ** Complex onset is a worse violation than DEP-V; therefore, si.te.?o.nu emerges as the optimal candidate, even though it violates DEP-V thrice. Coda Deletion e.g. otomati from English ?t?mætIk "automatic? Table 11: Coda Deletion STRPRES >>NO CODA >> IDENT-IO >> MAX-C STRPRES NO CODA IDENT-IO MAX-C ?.t?.mæ.tIk **! * ?.t?.mæ.tI **! * * ? o.to.ma.ti * * o.to.ma.tik *! * The relevant higher ranked constraints, namely STRPRES, and NO CODA, are violated by the other three candidates. This strikes out the other candidates and o.to.ma.ti emerges as the winning candidate. Cluster Break-up at the Onset e.g. buro?i from English br?? "brush? Table 12: Cluster Break-up at the Onset STRPRES >> COMP ONS >> NO CODA>> IDENT: IO >> DEP-V STRPRES COMP ONS NO CODA IDENT- IO DEP-V br?? *! * * ?bu.ro.?i * ** bu.ro? *! * * bro.? *! * * buro?i, the winning candidate, satisfies the highly ranked constraints - STRPRES, COMP ONS, and NO CODA. This results in its selection as the optimal candidate. Cluster Break-up at the Coda e.g. tebulu from English teIbl "table? Table 13: Cluster Break-up at the Coda STRPRES>> COMP CODA >> NO CODA >> IDENT: IO >> DEP-V STRPRES COMP CODA NO CODA IDENT- IO DEP-V teI.bl *! * * te.bl *! * * ?te.bu.lu * ** te.bul *! * * * te.bu.lu is the right output candidate. Structure Preservation e.g. d?uelu from English d?oel "personal name? Table 14: Structure Preservation STRPRES >> NO CODA >> ONSET >> IDENT-IO >> DEP-V STRPRES NO CODA ONSET IDENT- IO DEP-V d?oel *! * d?o.el *! * * ?d?u.e.lu * * * d?u.el *! * STRPRES and NO CODA are higher ranked constraints in Ma da . Therefore, their violations by the three other candidates make d?u.e.lu the optimal candidate. Summary of Findings and Directions for Future Research Summary This paper has shown that Ma da has borrowed extensively from Hausa and English. In the process of nativizing the loanwords, different phonological processes such as deletion and cluster break-up by insertion or deletion are involved. The paper has used the OT universal constraints to account for these various nativisation processes. The ranking of the constraints, which is language specific, is done according to the natural phonology of Ma da . The constraints that are ranked highest are the ones that apply to Ma da phonology, while those that are ranked lowest are the ones that do not apply to Ma da phonology. Based on the Ma da ranking of the constraints, the nativisation processes such as exact borrowings, initial vowel deletion, vowel insertion, coda deletion, cluster break-up at the onset, cluster break- up at the coda and structure preservation are well accounted for. Unresolved Issues There are certain areas in Ma da loanwords which this study cannot claim to have given adequate account of. For instance, voicing deletion is observed in the word kanga, which is borrowed from Hausa ganga "barrel-drum?. This adaptation violates the faithfulness constraint - MAX VOI. Similarly, the word b?ki, which is borrowed from English p?kIt "pocket?, violates another faithfulness constraint - DEP-VOI by adopting the voiced counterpart of the voiceless bilabial plosive /p/. However, these words and others have not been dealt with in this study. Instead, they have been left for future research. Secondly, there are instances of segment substitution as shown in Table 15. Table 15: Some Instances of Segment Substitution Substitute s Ma da form Source form Donor language English gloss d? < z d??gulu zogal Hausa horse reddish tree l < r l? ga riga Hausa shirt l? gu rogo Hausa cassava p < f p?lap?lu filafili Hausa oar Similar instances of segment substitution mentioned above were pointed out in Golston and Yang (2001), but theirs was a case of lack of those segments in White Hmong?s phonology. Ma da ?s case is quite different because these substituted segments (z, r, f) exist in Ma da . It is therefore not clear why these segments should be substituted in Ma da . Thirdly, the assignment of tone to Ma da loans from English pays no attention to the location of the stressed syllable in the English words. Tone loans from tone languages (Shugbu, Hausa, and Yoruba) were retained. It is hoped that models of tonogenesis will help to account for the tones found in loanwords. Fourthly, OT constraints have no way of accounting for syllable deletion in Ma da loanwords. The deletion can be word initial, as in: rube (Hausa) be (Ma da ), word medial, as in: lahadi (Hausa) ladi (Ma da ) or word final, as in: dama (Hausa) da (Ma da ). The source forms voilate no constraint in Ma da loanword phonology, yet syllable deletion occurs. It is not clear why syllables are deleted in such forms. This study aligns with Golston and Yang?s (2001:26-27) assertion that "[w]e are left with the impression that loanword phonology still poses formidable challenges to models of grammar. Although constraint-based analyses do substantially better at modeling loanword phonology in some areas (syllables and segments), they fail to provide much insight in others (such as tones)." However, this study whets an appetite for future research. References Akinlabi, A. (2008). Yoruba: A Phonological Grammar. Book Proposal Presented at the 39th Annual Conference of African Linguistics, Georgia, USA. Aziza, R. & Utulu, D. (2003). Loanword Phonology: English in Urhobo and Yoruba.Paper Presented at the 4th WOCAL, Rutgers University, New Jersey. Burenhult, N. (2001). Loanword Phonology in Jahai. Lund University, Dept. of Linguistics Working Papers, 48: 5-14. Golston, C. & Yang, P. (2001). White Hmong Loanword Phonology. Proceedings of HILP5. Haspelmath, M. & Tadmor, U. (eds.). (2009). Loanwords in the World's Languages: A Comparative Handbook. Berlin: Mouton de Gruyter. Kenstowicz, M. (2005). The Phonetics and Phonology of Korean Loanword Adaptation. Paper Presented at the First European Conference on Korean Linguistics, Leiden University. Kiparsky, P. (1982). Word-formation and the Lexicon. In F. Ingemann (ed.). Proceedings of the Mid-America Linguistics Conference. Lawrence, Kansas. McCarthy, J. (2002). A Thematic Guide to Optimality Theory. Research Surveys in Linguistics. Cambridge: Cambridge University Press. McCarthy, J. (2008). Doing Optimality Theory. Oxford: Blackwell. McCarthy, J. & Prince, A. (1993). Prosodic Morphology I: Constraint Interaction and Satisfaction. Report No. Ru CCS-TR-3. New Brunswick, NJ: Rutgers University Centre for Cognitive Science. McCarthy, J. & Prince, A. (1994). The Emergence of the Unmarked: Optimality Prosodic Morphology. In M. Gonzalez (ed.). Proceedings of the North East inguistic Society, 24: 333-379. McCarthy, J. & Prince, A. (1995). Faithfulness and Reduplicative Identity. In niversity of Massachusetts Occasional Papers in Linguistics, 18: 249-384. Oyebade, F. O. (1998). A Course in Phonology. Ijebu-Ode: Shebiotimo Publications. Oyebade, F. O. (2006). Prosodic Structure Preservation and English Loanwords in Yoruba: A Constraint-based Account. Paper Presented at the 20th Conference of the Linguistic Association of Nigeria, Abuja. Price, N. (1989). Notes on Ma da Phonology. The Summer Institute of Linguistics, Inc. Prince, A. & Smolensky, P. (1993). Optimality Theory: Constraint Interaction and Satisfaction. Report No. Ru CCS-TR-2. New Brunswick, NJ: Rutgers University Centre for Cognitive Science. Williamson, K. & Blench, R. M. (2000). Niger - Congo. In B. Heine & D. Nurse (eds.) African Languages: An Introduction. Cambridge: Cambridge University Press. Yip, M. (2002). Perceptual Influences in Cantonese Loanword Phonology. Journal of the Phonetic Society of Japan, 6: 4-21.