




































      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
1 

 

AN OPTIMALITY THEORY ANALYSIS OF NŊWƐ SYLLABLE STRUCTURE 

 

 

Esther Phubon Chie Epse Asonganyi  

University of Buea, Buea, Cameroon 

 

 

Introduction 

In the early forms of generative phonology, the syllable played no role (Clements and 

Keyser, 1983). Syllables are often considered as the phonological “building blocks” of words. 

They can influence the rhythm of a language, its prosody, its poetic meter, or its stress 

pattern. It is also considered “the smallest unit of speech that normally occurs in isolation, 

consisting of either a vowel alone (as in the pronunciation of the pronoun oar, ore, awe, I, 

eye, air, heir, err, our, hour, or, ah, etc.) or a combination of vowels and consonants (as in the 

pronunciation in, of so, it, two, four, salt, road, etc.)”. Some consonants can be pronounced 

alone and may or may not be regarded as syllables (syllabic nasals in most African 

languages), but they normally accompany vowels, which tend to occupy the central position 

in a syllable. 

  According to Clements and Keyser (1983), the primary set of core syllable types, 

cross linguistically contain the following sequences: 

(1)  a. V (Some languages allow V-initial syllables e.g. V or VC). 

   b. CV (Every language allows CV syllables i.e. no language bans C-initial syllable and   

on the other hand, no language equally bans V-final syllables.   

     c. CVC (Some languages allow final consonants e.g. VC or CVC). It is worth noting here 

that these are patterns of output syllable inventories but not restrictions on both 

morpheme and input shapes. 

Languages of the world have preferred syllable structures. Traditional work on internal 

structure of the syllable has arrived at the conclusion that the syllable is a phonological 

constituent which consists of zero or more consonants followed by a vowel and ending with a 

shorter string of zero or more consonants. The importance of these three spans has long been 

recognized and various names have been given to sub-parts of the syllable. They are 

generally referred to as the onset, the nucleus and the coda respectively as in the Nŋwɛ 

example below where the major parts of the syllable are indicated.   

(2)            

         /   \ 

       O     R 

       |      / \ 

       |   N   C 

       |    |     | 

       X    X   X  

        |      |     | 

       ŋ     á    ŋ      ‘bell’ 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
2 

 

The syllable is the prosodic category organising segments in sequences according to their 

sonority values. Each syllable has a sonority peak (nucleus), usually a vowel, possibly 

surrounded on both sides by margin segments of lower sonority, usually consonants (onset and 

coda), (Kager, 1999: 91). To obtain sensible words, the different pertinent, distinctive, 

invented, and organised elements have to be combined.  

  The main objective of this paper is to investigate the potentiality of Optimality Theory 

(OT) and to see whether one can depend on it to generate syllable structures cross linguistically 

and apply that general perspective on a particular language like Nŋwɛ. Consequently, this 

study is divided into four sections. The first section of this paper contains an introduction and a 

general presentation of OT, taking into account its components (Lexicon, Generator and 

Evaluator) and also the constraints (Faithfulness and Markedness). The second section 

involves a presentation of the syllable, the syllable structure constraints and the ranking of 

these syllable structure constraints in the language. The third section on its part is concerned 

with the application of OT to the analyses of the syllable structures in the language. In this 

regard therefore, syllable constraints are applied using some Nŋwɛ data. Finally, the fourth 

section summarises the paper presenting the findings of the work.  

 

 

General presentation of optimality theory 

  Linguistics, just like every other science, has known successive developments in the 

course of its evaluation to the extensive fieldwork and fine-grained analysis of data from 

languages of different families. From this large body of research a broad picture emerges of 

‘unity in variety’: core properties of grammar (with respect to sub-system of sounds, words, 

phrases and meaning) in a set of universal properties. Grammars of individual languages draw 

their basic options from these limited sets, which many researchers identify as universal 

grammar (UG). Each language, thus, reflects in a particular way the structure of its specific 

characteristics.  

  OT, which was developed by Prince and Smolensky in 1993 presents a very different 

view of phonological studies and is now also used in the study of syntax and language 

acquisition. It is a Constraint-Based Approach as opposed to the Derivational Approach 

inaugurated by Chomsky and Halle (1968) in the Sound Pattern of English (SPE). The 

derivational approach normally takes the form “X becomes Y in the context of Z” (X→Y/-Z).  

  According to McCarthy (2002) “Optimality Theory is the circulation of grammatical 

well-formedness which is accomplished by the optimalization of a set of constraints on 

structures and on input-output disparity, instead of through serial application of rule subject 

and filtering constraints.” The fundamental ideas of optimality theory are, firstly, that there 

are no rules; everything is done by constraints.  Secondly, that all constraints are allowed to 

be violated. It is not the violation as such that fails a candidate’s output but the ranking. If one 

constraint is more important, and more highly ranked than another, then any candidate that 

violates that one is rejected and the remaining candidate may be acceptable even if it goes on 

to violate the lower ranked constraints. 

 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
3 

 

 In optimality theory markedness is built into grammar in the form of universal output 

constraints, which directly state marked or unmarked patterns. In optimality theoretical terms, 

constraints are violable. The Violation of a constraint is not a direct cause of ungrammaticality, 

nor, the absolute satisfaction of all constraints is what depicts the optimal candidate. 

Constraints are intrinsically in conflict, and every logical possible output of any grammar will 

necessarily violate at least some constraints. Languages differ in their ranking of these 

constraints. Each violation of a constraint is avoidable, yet the violation of a higher ranked 

constraint is avoided more forcefully than the violation of a lower ranked one. 

  At the heart of OT lies the idea that language in general and every grammar in 

particular is a system of conflicting forces. These forces are embodied by several constraints, 

each of which makes a requirement about some aspects of the grammatical output forms. 

While constraints are universal, the ranking is not; differences in ranking are the source of 

cross-linguistic variation. For a given input, the grammar generates and later evaluates a set of 

output candidates, from which the optimal candidate is selected, which is the actual output. 

Evaluation takes place by ranking the constraints starting with the higher ranked before 

moving to the lower ranked ones, each of which will eliminate some candidate’s output, until a 

point where only one output survives. 

  The optimal output candidate is one which is harmonic with respect to the set of ranked 

constraints, that is the one that violates the least constraint(s). Harmony is a kind of relative 

well-formedness condition taking into account the severity of the violation of individual 

constraints as determined by their hierarchical ranking. OT allows ranking of variation present 

in human languages and it also constraints that which is not possible in some languages.  

 

 

Components of optimality theory grammar 

 

LEXICON: Contains lexical underlying forms of morphemes, which form the input. 

GENERATOR: Generates output candidates for some inputs, and submits them to the 

evaluator. 

 

EVALUATOR: Evaluates the different candidates that have been generated by the generator 

and selects the optimal candidate. The roles of GEN, EVAL, and CON are illustrated 

schematically below using the Nŋwɛ word: /è-gwíd/ ‘maggot’. 

 

 

 

 

 

 

 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
4 

 

3)  input:    /è-gwíd/                      GEN 

 

 

 

 

 

 

 

candidate set:           è-gwíd   è-gwídí    è-gwí      gwíd è-wídí  

 

 

 

 

 
 

                                                                   EVAL 

                                                              (constraints) 

  
 Optimal output:       [è-gwíd] ‘maggot’ 

   

 

From the diagram above, we see that /è-gwíd/ ‘pepper,’ is used as the input. From the input, 

the GEN creates many candidates, namely, /è-gwíd/, /è-gwídí/, /è-gwí/, /gwíd/ and even /è-

wídí/. Note that GEN can create many more candidates from this input other than the ones 

mentioned.  EVAL on its part uses the different constraints (Faithfulness – Universal 

constraints- and Markedness- language specific-) to select the optimal candidate. This theory 

is different from the others because it shows that constraints are conflicting and the primary 

action in OT is comparative.  

 

Constraints in OT 

McCarthy (2002) identifies two main constraint types: faithfulness and markedness, some of 

which are used in this paper.   

 

Faithfulness constraints 

Faithfulness constraints ban disparity between input and output forms as follows: 

a)   MAX-IO:   Input segments must have output correspondents (‘No deletion’).  

b)   DEP-IO:   Output segments must have input correspondents (‘No epenthesis).  

  They require that output forms preserve the properties of their basic (lexical) forms, 

expecting some kind of similarity between the output and its input.  

  

 

 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
5 

 

Markedness constraints 

Markedness constraints are language specific. They require that output forms meet some 

criterion of structural well-formedness. They take the form of prohibitions of marked 

phonological structures. Examples of markedness constraints relevant to Nŋwɛ are: 

a. Vowels must not be nasal. 

b. Syllables must not have codas.  

c. Syllables must have onset. 

d. Complex onsets are not allowed. 

e. Complex codas are not allowed.        

The relevant syllable markedness constraints in Nŋwɛ used in this paper are presented 

below. According to Kager (1999) syllable constraints take the following format: 

 

a) ONSET   *[V   (‘Syllables must have onsets.’) 

b) NO-CODA   *C]   (‘Syllables are open.’) 

c) *COMPLEXONS  *[ CC       (‘Onsets are simple’) 

d)      *COMPLEXCOD *CC]       (‘Codas are simple’)  

 

Nŋwɛ syllable structure constraints  

The ranking of Nŋwɛ syllable structure constraints  

  The basic syllable structure constraints, MAX, DEP, ONSET, and NO-CODA, are 

ranked here in dominance order suitable for Nŋwɛ. Two points which are crucial here are: 

whether onsets are required and/or whether codas are forbidden. Also, if either is, how is that 

enforced? The answer to these questions will determine the ranking of markedness 

constraints with respect to their faithfulness counterparts. Attention will be paid first to the 

question that is related to Onset and Codas. Not all the syllable types in the language have an 

onset, which means that onset is optional in Nŋwɛ. This goes to illustrate the fact that 

ONSET is dominated by the faithfulness constraints. This relation of dominance is 

represented as DEP>>MAX>>ONSET. 

  On the other hand, not all the syllable types in the language have codas, pointing to 

the fact that codas are optional in the language thus, not forbidden. This and in accordance 

with what has been discussed, results in ranking NO-CODA lower than the faithfulness 

constraints as follows: DEP>>MAX>>NO-CODA 

  This means that the relative order of the markedness constraints on the one hand and 

the faithfulness constraint on the other in Nŋwɛ shows that the faithfulness constraints 

dominate the markedness ones and this can be represented in as 

DEP>>MAX>>ONSET>>NO-CODA.  

It is necessary to examine this ranking to see whether it succeeds in determining the 

true output if it were to evaluate particular inputs. First, I take as the input CVC and try to 

examine the best candidate as shown in the analysis in the following table (a table on which 

constraints are ranked horizontally in a descending ranking from left to right, and output 

candidates are ranked vertically in a random order. The cells on the different tables contain 

violation marks ‘*’ incurred by each candidate for the constraint heading the column and the 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
6 

 

optimal candidate is marked by the index .The dotted lines separating two constraints 

indicate that the two constraints have no particular ranking i.e., either of the two can be 

ranked first (Kager, 1999: 3).  

In this table markedness constraints are ranked over faithfulness constraints.  

 

(4) Input:   /CVC /                                    output:  *[CV] 
 

     /CVC/     ONSET     NO-CODA   DEP MAX 

a)  CVC             *!                      

 b)  CV.CV                 *         

c)  CV                   * 

     Table 1 

 

It is shown in the table that ranking markedness constraints over faithfulness constraints 

results in the wrong output. Candidate (c) which is the CV syllable type emerges as the 

optimal candidate whereas (a) is supposed to be the optimal candidate. If faithfulness 

constraints are ranked over markedness constraints we will have the right output. 

 

 

(5) Input:    /CVC /                                 output:  [CVC] 
 

     /CVC/     DEP     MAX   ONSET   NO-CODA 

a)   CVC            * 

b)     CV.CV       *!    

c)       CV          *!   

     Table 2 

   

In the table, the constraints are ranked in a reverse order and the correct output is obtained. 

Obviously, the postulated ranking of constraints is successful in determining the true output 

of the input CVC, since candidate (a) emerges as the optimal candidate.  

 

Nŋwɛ syllable structures  

  Nŋwɛ canonical syllable types include V, CV, CVC, and N while the derived 

canonical syllable structures are CGV, NCV, NCVC, NCCV, and NCGV. Other syllable 

types can be seen to occur in a combinatory manner in the language.  

  An observation about the syllable types in this language is that not all the syllables 

begin with a consonant, which means that onset is optional in the language; therefore there 

are ‘onsetless syllables’.      

  

Analysis of Nŋwɛ syllable structures 

The V syllable structure 

  Here, the ONSET constraint is ranked with respect to faithfulness constraints in 

order to determine whether onsets are required or not. On the other hand, faithfulness 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
7 

 

constraints are ranked with respect to each other so as to show the ONSET requirement in 

Nŋwɛ. To do so, let me take the input /V/ and evaluate a set of three candidates, each of 

which exhibits a single violation of one of the four constraints in question, against different 

constraint hierarchies showing the possible relative ranking.  

a) V     violates     ONSET 

b) (V)   violates     DEP 

c)        Vv    violates     MAX                         

  When V is taken as the optimal candidate, it means that the constraints, which are 

violated by V, must appear very low in the constraint hierarchy related to the syllable structure. 

Therefore, in order to get the optimal candidate /V/ in Nŋwɛ, faithfulness dominates 

markedness. This means that constraints under faithfulness dominate those on markedness to 

get the optimal candidate in this language as said above. Consider the example and table 

below. 

 

(6)   /óÎ/        ‘mother’ 

      /m ̀/          ‘yes’ 

  Markedness constraints will still be ranked over faithfulness constraints here so that 

we see whether the optimal output will be derived. 

 

Input: /óÎ/  ‘mother’                                                          output:  * [bóÎ] 
 

       /óÎ/ NO-CODA       ONSET      DEP    MAX 

a)    óÎ                *!                 

b)    bóÎ                   *        

   c)    óÎg        *!           *         *          

    Table 3  

 

 Input: /óÎ/    ‘mother’                                                 output:   [óÎ] 
 

       /óÎ/      DEP      MAX     NO-CODA    ONSET 

a)   óÎ                   * 

   b)    bóÎ           *!    

   c)    óÎg           *!          *         * 

    Table 4 

   

Considering tables 3 and 4, the correct ranking would be to have faithfulness constraints 

outrank markedness constraints as earlier seen above in the CVC analysis in tables 1 and 2. 

This is so because if the ranking is reversed, candidate (b) will emerge as the optimal 

candidate being a CV structure and not a V structure which is the input. It is realised that 

onset is not required therefore making it possible for ONSET and NO-CODA to be lowly 

ranked, permitting DEP and MAX to dominate them in order to get the optimal output. 

Therefore, the constraints are ranked in the order DEP>>MAX>>NO- CODA>>ONSET. 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
8 

 

  The V syllable type can combine with other syllable types to form words in the 

language. Therefore, I consider the constraints and the ranking that derive these combinations 

as follows: 

  

V.CV combination   

  In this combination the first syllable contains a vowel and the second is made up of 

both an onset and a nucleus. Consider the examples and table below. 

 

(7)     à.bè ‘profit’ 

         e ̀.fó̀          ‘father’ 

         à.bo         ‘hand’ 

         e ̀.sa ̄         ‘market’ 

        a ̀.lí  ‘blood’ 

 

   input:  /à.bè / ‘profit’                                     output:   [à.bè] 
 

  / à.bè/  DEP  MAX    ONSET    NO-CODA 

a)  à. bè               *  

 b)  à.bèt      *!              *          * 

 c)   à.b     *!             *  

    Table 5  

 

It is assumed in this paper that the right ordering for the constraints would be faithfulness 

over markedness because if the reverse is taken as seen in the first and third tables above, the 

wrong output is gotten. There would be no need for the comparison of the ranking of the 

constraints since the right output is already obtained. Therefore, in table (5), the GEN 

produces candidates from the input which violate ONSET and NOCODA. This does not 

disqualify all the candidates, since these are the lower/lowest ranked constraints in the 

language. Candidate (a) is the optimal candidate and this is due to the fact that it satisfies the 

highest ranked constraint thereby violating the lowest ranked constraints as seen above. On 

the other hand candidates (b) and (c) violate DEP and MAX which are the two highly ranked 

constraints thereby disqualifying them. 

 

The V.CVC combination 

  A similar ranking hierarchy will be used at this point to see if it derives this syllable 

structure present in the language. Examples are given below.  

(8)     è.sáŋ   ‘broom’ 

          e ̀.lùɁ   ‘honey’ 

          a.tɛ́m   ‘hut’ 

          a ̄.fɛ́m        ‘breeze’ 

          e ̀.bu ̀Ɂ         ‘mushroom’  

     input: / e ̀.lùɁ /   ‘honey’                                   output:     [e ̀.lu ̀Ɂ] 
 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
9 

 

/ e ̀.lùɁ/  DEP     MAX    ONSET  NO-CODA 

a) le ̀.lùɁ     *!                  * 

b)   è.lùɁ                    *         * 

       c)    è.lù         *!          *  

    Table 6 

   

The optimal candidate (b) satisfies all the faithfulness constraints and equally violates all the 

markedness ones, which are lowly ranked. Here the constraint hierarchy is maintained since it 

is what is used in Nŋwɛ. When it comes to determining the optimal candidates in syllable 

structures, stating the right ranking of the constraints is important and needs to be handled 

with great care because the wrong ranking change will give priority to the wrong candidate 

and violate the correct syllable structures of the language. Consider another combination 

under the V syllable type, which is an open one.  

  

The V.CV.CV combination 

  This is also a complex syllable structure combination and it is mostly present in 

trisyllabic words, which are numerous in Nŋwɛ verbs. The centre syllable is closed but the 

beginning and the final syllables are open. Examples are given below. 

(9)        a.dì.dì               ‘chin’ 

             a ̀.kó́.ró ̀              ‘waist’ 

     à.ʒó ̀.gó ́      ‘afternoon’ 

     à.la ́.ku ̀      ‘foot’ 

             e ̀.kì.bì    ‘plum’ 

    

    /input: / a.dì.dì/  ‘chin’                                 output:  [a.dì.dì] 

  / a.dì.dì/ DEP MAX     NO-CODA      ONSET 

 a)   a.dì.dìɁ   *!                *              * 

    b)     a.dì.d     *!                 *        

   c)    a.dì.dì                             * 

     Table 7 

   

In the table above, candidates (a) and (b) are eliminated because they violate the top-most 

ranked constraints DEP and MAX. These two candidates violate them in order to satisfy the 

lower constraints ONSET and NO-CODA, and this is not accepted in Nŋwɛ, since onsets and 

codas are optional elements in the language, leaving candidate (c) to emerge as the winner.  

  Interestingly, one would expect to have the V.CVC.CV syllable type in Nŋwɛ. 

However, this is not the case because, VCV combination can only syllabify to V.CV and not 

VC.V and the CVCV combination can also only syllabify to CV.CV and never CVC.V.   

 

 

 

 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
10 

 

The V.CV.CVC combination 

  The V.CV.CVC combination in Nŋwɛ is made up of three different canonical syllable 

structures which are V, CV and CVC which have already been discussed above. The analysis 

of this syllable combination combines the three different syllable structures in the language. 

Examples include the following.   

(10)      a ̀.té.sék             ‘mat’ 

             a ̀.kí.mó́k           ‘charcoal’ 

             a ̀.té.tà̀ŋ              ‘Irish potatoes’ 

       

      input: /a ̀.té.ta ̀̀ŋ/ ‘Irish potatoes’                       output:  [a ̀.té.tà̀ŋ]   

     /à.té.tà̀ŋ/     DEP MAX      ONSET  NO-CODA 

   a) à.té.ta ̀̀ŋ          *        * 

b)    à.té.ta ̀̀ŋ.é            *!         **         * 

c)  a ̀.te ́.ta         *          **               

        Table 8 

   

From the table above it is realised that candidate (a) is the optimal candidate because it violates the least 

ranked constraints. The first syllables do not have onsets and do not have codas. Nothing is inserted or 

deleted from the syllables i.e., input is exactly like output. Candidate (b) violates DEP once and violates 

ONSET and NO-CODA twice because of the vowel /é/ which is inserted to create a new syllable. 

Candidate (c) on its part violates MAX and ONSET but doesn’t violate NO-CODA because the second 

and third syllables do not have codas.  

 

The CV syllable structure 

  This is the syllable type which is considered to be universal. No language prohibits 

the CV syllable structure, thus no language prohibits onsets or require codas. This means that 

these languages have ‘Onsetful syllables’, open syllables, and may allow codas. Hence, 

languages have optional consonant initial syllables but never ban them and optional vowel 

final ones but never require them.  

  Apparently, there are two operations according to Clements and Keyser (1983) 

which are responsible for producing the three less natural core syllables from the most natural 

one, CV.  The first involves the deletion of the initial consonant of the CV syllable type to have a V, 

or it can as well be considered as being the first of the two steps towards producing VC. The second 

one involves adding a final consonant to the CV syllable type to have CVC. This predicts the different 

operations to be used to get the CV structure in Nŋwɛ. Consider the data and table below. 

 

(11)    ŋó               ‘snake 

          ŋó                ‘sun’ 

          be                 ‘pick’  

          teà                  ‘not’ 

          tʃě                ‘urinate 

        

       input:  / tʃě /  ‘urinate’                                      output:  [tʃě] 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
11 

 

 

       / tʃě/     DEP       MAX      ONSET NO-CODA 

   a) tʃě     

b) tʃěb         *!            * 

      c) ě               *   

      Table 9 

   

The CV syllable structure is universally optimal because it doesn’t violate any of the syllable structure 

constraints. No alternative analysis, therefore, can be more adequate than what is presented. The basic 

CV syllable theory constraints are: ONSET, NO-CODA, NUC (nucleus), *COMPLEX, DEP, MAX 

etc. This is obviously due to the simple fact that this syllable type doesn’t violate any of the universal 

constraints. There are other syllable combinations, which fall under this syllable type, and they are 

presented below. 

 

The CV.CV combination 

        It is a combination with no coda and it has its nucleus in the final positions of each of the 

syllables and an onset. This is therefore analysed as shown below.  

(12)         lefó          ‘eight’ 

                tó́Ɂó́          ‘tomorrow’ 

                beró        ‘nose’  

                lɛ́dá            ‘write’ 

                óŋa ́          ‘suck’ 

             

     input: / lɛ́da ́/ ‘write’                                    output:   [lɛ́dá] 
 

        / lɛ́dá /                  DEP                            MAX      ONSET   NO-CODA 

    a)   lɛ́.á         *!          *         

 b)  lɛ́dá                                

    c) lɛ́d.áp   *!                   *         * * 

                   Table 10 

   

In table (10) above, the optimal candidate (b) satisfies all the constraints, thereby evolving as the 

optimal candidate, while, candidate (a) violates MAX and ONSET due to the deletion of the 

consonant /t/ in the onset position of the second syllable. The last candidate violates DEP, ONSET 

and NO-CODA due to the transfer of the consonant /d/ from the onset position of the second syllable 

to the coda position of the first syllable, thereby, leaving the second syllable with no onset. It also 

violates NO-CODA twice because of /t/ and /p/ at the coda positions of the two syllables. 

The CV.CVC combination 

       This combination is mostly present in disyllabic and polysyllabic words in the language. 

Consider the examples and the analysis in the table below. 

(13)         lèke ̄ŋ             ‘pot’ 

               be ̀ve ́d             ‘oil’  

               bɛ́bɛ́g             ‘lizard’ 

               lèle ̄Ɂ                ‘yam’ 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
12 

 

               lèfók                ‘indian bamboo’ 

     

       input: /le ̀ke ̄ŋ/ ‘pot’                           output: [le ̀ke ̄ŋ]  
 

        /lèkēŋ/          DEP MAX ONSET  NO-CODA 

     a)   lè.kē     *!   

      b)   lès.ke ̄ŋ    *!          ** 

   c) lèke ̄ŋ                       * 

Table 11 

   

In table (11) above the first candidate violates MAX because of the deletion of the consonant 

/ŋ/ at the coda position of the second syllable. Candidate (b) is eliminated due to its violation 

of DEP and NO-CODA despite the fact that it satisfies the highly ranked MAX constraints. 

The optimal candidate is (c) because it violates the lowest ranked constraint. 

 

The CV.CV.CV combination 

      With the CV.CV.CV combination, it is realised that no matter the length of the syllable, it 

is still an open syllable. It is also universally optimal despite the fact that it has three different 

syllables which are all CV. The analysis of this structure is similar to that of the CV.CV 

syllable structure in table (11) above. This is because they have certain characteristics in 

common. These characteristics include the following: They are open, hence, have no codas and 

they all have onsets. It is worth noting that this syllable type is so rare in the language. 

Examples include the following.   

(14)      letoŋe              ‘ear’ 

             lègódʒó ̀          ‘now’ 

             lèbróɁeà       ‘nine’ 

     bèɳaróà        ‘sand’ 

     ŋáɁáse ́             ‘bend down’         

  

        input: /be ̀ɳaróà/ ‘sand’                       output:  [be ̀ɳaróà]   

     /bèɳaróà/     DEP MAX ONSET  NO-CODA 

   a) bè.ɳa.róà     

b)    bè.ɳa.róàp            *!               * 

c)  be ̀ɳ.a.róà                       *              * 

        Table 12 

From the table above it is realised that candidate (a) is the optimal candidate because it doesn’t 

violate any of the constraints. All the syllables have onsets and do not have codas. Nothing is 

inserted or deleted from the syllables i.e., input is exactly like output. Candidate (b) violates 

DEP and NO-CODA because of the consonant /p/ which is inserted in the coda position of the 

last syllable. Candidate (c) on its part violates ONSET and NO-CODA because the second 

syllable doesn’t have an onset and the first syllable has a coda. 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
13 

 

The CVC syllable structure 

     The focus here is to determine whether it is banned or optional. To do this, I have to take 

the input /CVC/ and evaluate it with a set of three candidates and see their exhibition of the 

violation of the constraints. Consider the analysis below. 

 

a) CVC            violates        NO-CODA 

b)         CV.CV        violates       DEP 

c) CV.(C)        violates        MAX 

 

  It is realised in this case that CVC which is the optimal candidate violates NO-CODA 

and satisfies ONSET. Therefore, since it is the optimal candidate, the constraint violated by it 

must appear very low in the constraint hierarchy as follows: DEP>> MAX>>NO-CODA. 

Therefore, the constraint hierarchy that is to be used is DEP>>MAX>>ONSET>>NO-CODA.  

  Again when both faithfulness constraints are highly ranked on the table, the candidate 

representing the very same input is designated the winner, even if such a candidate violates the 

other constraints, which in this case are the markedness constraints. See the data and table 

below. 

(15)      to ́ŋ             ‘basket’ 

             bu ̌k            ‘our’ 

             luàŋ             ‘hot’ 

             zuàŋ         ‘dry’ 

             ŋaàŋ          ‘bell’ 

        

    input: [luàŋ]   ‘hot’                           /su/  output:  [luàŋ]   

  /luàŋ/         DEP  MAX       ONSET   NO-CODA 

a)  luà            *           

 b)   luà.ŋ        *!            

   c) luàŋ                    * 

        Table 13  

   

When the same line of argument as that above is pursued, an inevitable conclusion is arrived 

at, i.e., DEP>>MAX>>Onset>> NO-CODA so, candidate (c) wins. Therefore, codas in Nŋwɛ 

are optional. It is also realised that if codas would have outranked the pairs of markedness 

constraints, then codas would have been forbidden in the language and the banning would have 

been reinforced by either deletion or epenthesis that is being decided on by the relative ranking 

of DEP and MAX. The banning of codas is not a functional element in Nŋwɛ due to the fact 

that they are optional, not forbidden characteristics.  

  This syllable type doesn’t really have combinations with other syllable types. It has a 

few words with CVC.CVC and CVC.CV combinations.  

 

 

 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
14 

 

Analysis of some derived syllable structures  

The derived syllable types in the language result from the modification of the 

canonical syllable types as a result of some phonological processes such as devocalization 

and prenasalization or both. It is worth noting here that the focus of this paper is not to 

discuss these phonological rules but to handle these derived syllable structures using 

Optimality Theory. Looking at the syllables that are derived here, it can be observed that 

some of them contain complex onsets which means that complex onset is optional in Nŋwɛ. 

This goes to illustrate the fact that *COMLPEX (onsets and codas are simple) is dominated 

by the faithfulness constraints. A structural well-formedness constraint is needed to capture 

the margin of complexity of these syllable types since they are made up of more than one 

consonant in the same environment. This relation of dominance can be represented as 

follows: DEP>>MAX>>ONSET>>NO-CODA>> *COMLPEX. 

  From the above ranking, it is seen that *COMLPEX is the least ranked of all the 

constraints; therefore, all the other higher ranked constraints take precedence over this 

constraint. Worthy of note is the fact that all the markedness constraints here can be 

interchanged but we shall still have the same result. Some of the derived syllable structures 

will be analysed here. 

 

 

CGV syllable structure 

  This syllable type is made up of a consonant, a glide and a vowel. Glides are 

considered as consonants in this work since they occur on the consonant chart of the 

language. This syllable type exists mostly with verbs in the language.  Examples include the 

following. 

(16)  /fie/  [fjə́]  ‘give’ 

        /kui/  [kwí]  ‘receive’ 

        /dió/   [djó́]  ‘say’ 

        /zió /  [zjó ́]  ‘plant’ 

        /gui/  [gwĭ]  ‘laugh’       

 

   input:  /kui/  ‘receive’                                         output:  [kwí] 

     / kui/ DEP MAX ONSET NO-CODA *COMLPEX 

   a) kwí                * 

      b)   kwít  *!     *            * 

      c)   kw  *!                  * 

 Table 14  

From the table above, the optimal candidate is (a) because it violates only *COMLPEX which 

is the least ranked of all the constraints. Candidate (b) violates *COMLPEX, DEP and NO-

CODA because /t/ is added in the final position of the syllable thereby violating a no addition 

constraint, as well as codaless and simple codas constraints. Candidate (c) on its part violates 

both *COMLPEX and MAX because the final vowel /í/ is deleted and the candidate is made 

up of two consonants.   

 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
15 

 

NCV syllable structure   

Examples include the following: 

 (17)         ǹʒì   ‘sheep’ 

                 m ̀bú                      ‘dog’ 

                 m̀bɛ́  ‘flute’ 

                 ǹtá                        ‘thigh’ 

                    n ̀tɛ́                          ‘five’       

                

   input: /Ntɛ́/  ‘koki beans’                                             output:   [n ̀tɛ́] 

     /ǹtɛ́/       DEP      MAX       ONSET  NO-CODA 

a)  tɛ́         *!   

   b)    ǹ.tɛ́p         *!               * 

  c)   ǹ.tɛ́                                      

Table 15 

From the table above it is realised that candidate (c) is the optimal candidate because it doesn’t 

violate any of the constraints. The two syllables have onsets and do not have codas. Nothing is 

inserted or deleted from the syllables i.e., input is exactly like output. Candidate (b) violates 

DEP and NO-CODA because of the consonant /p/ which is inserted in the coda position of the 

last syllable. Candidate (a) on its part violates only MAX because of the deletion of the 

syllabic nasal. Worthy of note is the fact that *COMLPEX constraint is not necessary in the 

table above since there are no complex onsets or codas. 

 

NCCV syllable structure 

Examples include the following. 

(18)     ŋ̀kró ̌         ‘handle’  

 ǹtrɛ̌   ‘distance’ 

 ǹdró̌         ‘dress’ 

 ǹkrū         ‘monkey’ 

 ŋ̀kru ̀         ‘rope’ 

   

            input:  /ŋ̀kró̌/   ‘handle’                                   output:    [ŋ̀kró ̌] 

     /ŋ̀kró ̌/ DEP MAX ONSET NO-CODA *COMLPEX 

   a) ŋ̀kró̌p      *!                  *        *** 

 b)  ŋ̀kró ̌                                     *** 

        c)  ŋ̀kr         *!                                  *** 

                 Table 16 

In table (16) above, candidate (b) emerges as the optimal candidate because it violates the 

least ranked constraint *COMPLEX. Candidate (a) violates DEP, NO-CODA and 

*COMLPEX. This is seen in the insertion of the consonant /p/ which violates DEP and forms 

the coda of the second syllable. Candidate (c) violates MAX, and *COMLPEX which equally 

disqualifies it to become the optimal candidate.  

 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
16 

 

NCGV syllable structure 

Examples include the following: 

(19)      /ŋ̀kuó ̌/  [ŋ̀kwó̌] ‘gun’ 

             /n ̀zió́/  [ǹzjó́] ‘palm nut’ 

      /n ̀ʒió́/        [n ̀ʒjó́] ‘mirror 

            / ŋ̀guí/        [ŋ̀gwí]        ‘wife’ 

            /m ̀biə̀/      [m ̀bjə̀]      ‘cricket’ 

          

   [m ̀bjə̀]   ‘cricket’ 

        input: /Nbjə̀/                                            output:  [m ̀bjə̀] 

     /m̀bjə̀/  DEP MAX ONSET NO-CODA *COMLPEX 

   a) m̀.bjə̀                * 

     b)    m̀.bjə̀k   *!       *            * 

      c )   m̀.bə̀     *!                                        

 Table 17 

    

From the table above, the optimal candidate is (a) even though it violates *COMPLEX. 

Candidate (b) violates DEP, NO-CODA and *COMPLEX because a /k/ sound is added in the 

final position of the syllable thereby violating a no insertion constraint and a codaless 

constraint. Candidate (c) on its part violates MAX because the consonant /w/ is deleted 

creating a simple onset. 

 

 

Conclusion 

       In conclusion, it can be said that OT shows high potentiality in serving as a theory for 

the analysis of the syllable structure of Nŋwɛ. Throughout this paper, attention has been on a 

set of more general issues in OT as far as the syllable is concerned. It has been shown that 

syllable-dependent processes like any other phonological processes, involve interactions of 

‘markedness and faithfulness constraints. A change (deletion, insertion) is made at the cost of 

some faithfulness constraint, but only to avoid the violation of higher ranked constraints.  

  When a look is given to the 17 tables of this paper, it is realised that all the candidates 

(a), (b), and (c) emerge as the optimal candidate at different times and different tables in the 

analyses. This is due to the manner in which the selected candidates are arranged and also 

due to the ranking of the constraints. It is therefore due to this ranking that the following are 

suggested as the constraint ranking orders for Nŋwɛ syllable structure. 

 

(A) DEP>>MAX>>ONSET, (B) DEP>>MAX>>NO-CODA, (C) DEP>>MAX>>ONSET>>NO-CODA  

(D) DEP>>MAX>>NO-CODA>>ONSET and finally; (E) DEP>>MAX>>ONSET>>NO-CODA>> 

*COMLPEX. 

  Since GEN generates all logically possible candidates for a given input, OT grammar 

needs no rewrite rules to map inputs onto outputs. The evaluation of these candidate analyses 

is the function of the evaluator, the component of ranked constraints. It has equally been 



      

Current Issues in Linguistics, Language and Gender Studies: A Festschrift in Honour of Professor Cecilia Amaoge Eme 

 
17 

 

noticed in this study that candidates generated by the generator must not look exactly like the 

input in order to arrive at the optimal output as illustrated in the last five tables. 

   The discussions in this paper reveal that in Nŋwɛ just like other languages, 

faithfulness constraints outrank the language specific counterparts as far as the syllable 

structures are concerned. 

 

 

 

References 

 

Chomsky, N. and M. Halle. 1968. The Sound Pattern of English. New York: Harper and 

Row. 

Clements, G. N and S. J. Keyser. 1983. CV Phonology: A generative Theory of the syllable. 

Cambridge, Mass: MIT Press. 

Kager, R. 1999. Rhythmic Vowel Deletion in Optimality Theory.  In Derivations and  

Constraints in Phonology. (ed.) Iggy Roca, pp. 463-99. New York: Oxford University 

Press. 

McCarthy J. 2002. A Thematic Guide to Optimality Theory. Cambridge: Cambridge 

University Press. 

Njika J. 2001. “The Internal Structure of Nŋwɛ Narratives: A Study in Discourse Analysis 

and Application.” Diss. University of Yaounde I.  

Nkengsong P. 2014. “Syllable Structure in Nŋwɛ”. Long Essay for the Award of a B.A 

Degree in Linguistics. University of Buea. 

Phubon, E. C. 2014.  The Syllable Structure of Kejom (Babanki). In Epasa Moto: A 

Multidisciplinary Journal of Arts, Letters and the Humanities of the University of 

Buea, pp 32-60. Vol. 32 no.4. 

Prince, A. and S. Smolensky. 1993. Optimality Theory: Constraint Interaction in Generative 

Grammar, RuCCs Technical Report. no. 2, Piscataway, N.J. Cambridge, 

Massachusetts: Blackwell.  


