Carol Anyagwa
Abstract
This paper is a report on the pilot study investigating the geographical/ethnic dimension to the variable word stress patterns of Nigerian English. Fifty polysyllabic English words were read by fifty Nigerian undergraduates of the University of Lagos of varied socio-economic, educational and ethno-linguistic backgrounds. Each half of this sample was made up of L1 speakers of Igbo and L1 speakers of Yoruba respectively, whose productions were analysed statistically and metrically. The data were analysed by converting the tokens of occurrence to percentages, the higher percentages being taken as the norm in each accent. The perceived norms were further captured using the arboreal and SWS notations of Metrical Phonology. In spite of earlier declarations that Nigerian English (NE hereafter) is a homogenous variety in terms of stress assignment, the results of the study highlight the differences in the stress patterns of Igbo English (IE hereafter) and Yoruba English (YE hereafter) – both regional accents of NE; leaving us with the puzzle of deciding between such variable stress patterns as PROtein~proTEIN, UMbrella~umBRELla, CEremony~ceREmony and EARring~earRING which should be adopted as the NE stress patterns of these words. The study, while emphasizing the need for the exploration of the geo-ethnic approach to the definition of NE, establishes that the differences in the phonologies of regional accents of NE exist not only at the segmental level but also at the supra-segmental level of stress.
Keywords: Stress, Metrical Phonology, Igbo English, Yoruba English, Standardisation
Introduction
Despite the relatively wide acceptance and global recognition accorded the Nigerian variety of English, it is surprising that NE scholars are still battling with its description; what with the discrepancies in their representation of the accent to the outside world. Since the 1970’s when studies on the NE accent began to gather momentum, scholars within and outside the country have consistently sought to describe the basic features which characterise this accent. The multiplicity of pronunciation patterns within NE, particularly those induced by the extreme linguistic diversity within our national boundaries, has not in any way made the task of describing the dialect in monolithic terms easy. Although a number of studies have come up with corpuses of words and their NE pronunciations, on a number of occasions, the validity of such data has been denied by fellow NE scholars. Simo Bobda (2007) reports instances of denial by NE phonologists of the occurrence of prep[a]re and [pɔ] (poor) – which were hitherto attested in earlier data available to him – in NE. This suggests a possible disagreement among NE scholars on what the specific NE pronunciations of certain words are.
A possible source of this disagreement is the reliance of some of these scholars on the different pronunciation patterns which dominate particular ethnic varieties. The delay in the codification process, which implies the absence of dictionaries embodying NE usage, among other things, has also consistently ensured the sustenance of the multi-ethnic influences observed in the NE accent. We must, however, emphasize the fact that national varieties (like NE) usually refer to the variety spoken by the majority of educated users and not by a particular geo-ethnic section of a country. This implies that any serious study on NE should not be strictly based on data from subjects from a particular section of the country. Although several earlier studies have attempted a description of the segmental peculiarities of some L1-induced accents of NE, the novelty of this paper derives from its focus on a suprasegmental feature – stress – with emphasis on its internal variability among local accents of English in Nigeria. The purpose is to draw attention to this relatively-neglected area without which any attempt at codifying the Standard Spoken Nigerian English will be meaningless.
Regional Accents of Nigerian English
The subject of variability within NE has often been discussed using linguistic, educational and ethnic parameters, among others. Of all these, Jowitt (1991:38) observes: ‘An obviously attractive parameter for determining varieties within NE is the ethnic criterion i.e. distinguishing the various kinds of English that result from the influence of mother tongue transfers.’ Earlier, studies on NE accents were not based on individual ethnic groups, but on the two polar regions of the country. Jibril (1979), for instance, identifies two accents – the Northern and Southern accents – and their phonemic contrasts as presented below:
Table 1: Phonemic Contrasts between the Northern and Southern Accents of NE
| RP | Northern Accent | Southern Accent |
| θ | s | t |
| ð | z | d |
| æ | æ/e | ɑ |
| ʌ | a | ɔ |
| з | ɑ | ɔ /e |
Lending weight to the Northern/Southern accent distinction, Jowitt (1991:71) asserts that the differences between Igbo English (IE) pronunciation and that of Yoruba English (YE) – both Southern accents – are narrower than the differences between Hausa English (HE) – the Northern accent – and either of YE and IE. He attributes the striking differences which still exist between HE and IE/YE to the use of qualified native speakers of English by the colonial administration to teach English in Northern Nigeria and the fact that while the Igbo and Yoruba languages belong to the West Benue Congo languages and therefore share greater linguistic affinity; Hausa belongs to the West Chadic languages. He therefore concludes that the closer the languages, the more likely they are to share linguistic similarities.
In a subsequent study, Jibril (1982) identifies geographical varieties of NE in line with the three major regions of the country, namely: the Eastern, Western and Northern regions. This entails the lumping of minority languages among the three major ethno-linguistic groups – Hausa (North), Igbo (East) and Yoruba (West) (Dadzie, 2004). Thus, Nigerian linguists have come to confidently refer to these varieties as Hausa English, Igbo English and Yoruba English respectively (Jowitt, 1991:71; Osuafor, 2002:26 and Awonusi, 2004). Odumuh (1987) in a corresponding ethno-linguistic study of variation in Nigerian English, identifies three dialects: EngHausa, Engligbo and Yoruba English.
Jowitt (1991:47f), characterizing these ethnic (regional) varieties, prefers the term ‘Popular Nigerian English’ (PNE) – a broad spectrum variety within which he identifies PNE (I), PNE (Y) and PNE (H) where I, Y and H stand for Igbo, Yoruba and Hausa respectively. He argues: ‘The usage of every Nigerian user is a mixture of standard forms and Popular Nigerian English forms, which are in turn composed of errors and variants.’ Thus, the identification of PNE (I), PNE (Y) and PNE (H) suggests that certain Popular Nigerian English forms can be linked to indigenous language influences or, better still, traced to specific language groups. His analysis of the phonemic contrasts among these accents is presented below:
| RP | PNE (I) | PNE (H) | PNE (Y) | RP | PNE (I) | PNE (H) | PNE (Y) | |
| i: | i | i: | i | θ | t/ṱ | s | t | |
| ɪ | i | i | i | ð | d | z | d | |
| e | e/ɛ | e:/e | ɛ | p | p | f | p | |
| æ | а | а | a | v | v | b | f | |
| ʌ | ɔ | а | ɔ | z | z | z | s | |
| ɑ: | а | ɑ: | a | |||||
| ɒ | ɔ | o | ɔ | |||||
| ɔ : | ɔ | o: | ɔ | |||||
| ʊ | u | u | u | |||||
| u: | u | u | u | |||||
| з: | ɔ /a/ ɛ | a: | ɔ /a/ ɛ | |||||
| ə | u/ɔ/ɪ/e/ɛ | ə/o | ɔ/ɛ/u |
Table 2: Phonemic Contrasts among PNE (I), PNE (Y) and PNE (H)
Jowitt explains that in vowels which have alternative realisations, the choice is basically determined by spelling (e.g. /з:/ is /ɛ/ when spelt ‘ear’ as in ‘earn’, / ɔ / when spelt ‘ur’ as in ‘nurse’ and /ɛ/ or /a/ when spelt ‘ir’ as in ‘first’ and ‘sir’ respectively in PNE (I).
Jibril (1982) however observes the following alternations in Hausa English consonants: /p/→ [p / f]; /f/→ [p/ Ф] – where the realisation of the sounds varies between formal and casual speeches Awonusi (2004) adds that occasionally, the fortis velar stop /k/ may be labialized in Hausa English – /k/→ [k / kw] – while the fortis palato-alveolar affricate /ʧ/ is often realized as the fricative [∫] in Yoruba English. He also reports that the bilabial glide is often devoiced into [ʍ] or [hw] in wh-words in Igbo English; a phenomenon which Jibril (1982:93) ascribes to Irish or Scottish influence and which he identifies as occurring variably in some speakers and categorically in others. Jowitt goes further to identify certain phonological processes that are peculiar to individual accents which include vowel nasalisation (PNE (Y)), pre-syllabic glottal stop epenthesis and consonant gemmination (PNE (H)), and pharyngealization (PNE (I)). Other studies on the phonological peculiarities of the regional accents of Nigerian English (Awonusi, 1987; Adetugbo, 2004; Olaniyi, 2013) seem to be in agreement with the foregoing.
On the suprasegmentals, Dadzie (2004) observes that the Southerners’ stress realisation is different from the Northerner’s. The Southerner, he says, gives equal value to the syllables in all words in a sentence whilst his Northern counterpart realizes stress along the known patterns of English. He adds that the northerner is ‘more alive to the use of relevant intonation patterns than his friend in the south’ – a claim also upheld in the findings of Tiffen’s (1974) study in which the Northern accent is rated higher on the international intelligibility scale than the Southern accent.
Understandably, works on accents of English (e.g. Wells, 1982) concentrate on the segmental differences between native varieties more than the prosodic; the assumption being that divergences in stress placement, for instance, occur mostly in the Outer Circle. Unfortunately, however, in an Outer Circle variety like NE, we still observe a dearth of reading materials on the prosodic/suprasegmental peculiarities of both the larger dialect (NE) and individual regional accents. A number of studies have been carried out on the suprasegmental phonology of NE, all of which seem to agree on the notion that this aspect “is the last hurdle which majority of Nigerian speakers of English as a second language never manage to cross” Banjo (1979). Notable among these studies are Kujore (1985), Atoye (1991), (2005) and Simo Bobda (2010) on stress; Bolinger (1981), Udofot (2003), Akinjobi (2004) on rhythm; Gut (2001), Gut and Milde (2002), Jowitt (2000) on intonation.
Among the few studies on the stress systems of individual local accents of English are Jolayemi (2008), Omachonu (2008), Anyagwa (2013), Osifeso (2017) and Anyagwa (2015). Jolayemi (2008) acoustically analyses data from Ebira, Igala and Nupe accents of NE. The study reveals that ‘the stress pattern of NE is a preponderance of tone structure of the Nigerian indigenous languages’. Omachonu (2008), exploring the NONFINALITY and UNEVEN-IAMB constraints of the Optimality Theory examines NE stress using speech samples from Igala speakers of English. In line with the Optimality bias of the study, the result shows UNEVEN-IAMB >> NONFINALITY i.e. UNEVEN-IAMB is a preferred constraint to NONFINALITY. Combining acoustic analysis with Metrical analysis, Anyagwa (2013) analyses the stress system of the Igbo accent of NE. Her analyses revealed the preponderance of the S(trong) syllables in this accent, which alters the stress isochrony of English.
Osifeso (2017), also exploring the optimality framework using the NONFINALITY, UNEVEN-IAMB, and ALL-FT-LEFT constraints analyses the speech of one hundred mesolectal Yoruba speakers of NE. The result shows that word stress among Yoruba speakers of NE is usually rightward. In addition, the study affirms that the most prominent constraint among Yoruba speakers of English is UNEVEN-IAMB which prefers the assignment of stress on Light-Heavy syllables to Light-Light or Heavy syllables. In a study which can be described as a paradigm shift, Anyagwa (2015), adopting a comparative approach, demonstrates that Yoruba English and Igbo English significantly differ in stress realisation. Using data collected from fifty University of Lagos undergraduates of Yoruba and Igbo extraction, and analysed statistically using the t-test, the study highlights the internal differences between the stress systems of the Yoruba and Igbo accents of NE. The study identifies regional variation in word stress patterns as another clog in the wheel of progress of Standard Nigerian English codification process.
A major gap revealed in our review of available literature above is the dearth of comparative studies on the stress systems of regional accents of Nigerian English, probably as a result of the claim that “these sub-varieties exhibit a greater unity in their prosodic traits in terms of intonation and word stress than in their segmental features” Atoye (2005). This area thus remains one of the aspects of NE phonology crying for attention. Although Anyagwa’s (2015) study took the comparative dimension, this present study is distinguished by its application of the Metrical Theory of stress to the analysis of data from the two southern accents of NE. We submit that the extent of the supposed ‘greater unity’ which exists among the prosodies of these accents of NE can only be ascertained through empirical investigations of the individual accents’ prosodies.
Theoretical Framework
The Metrical theory of stress constitutes the major theoretical underpinning for the analysis of the data in this study. The theory has its origin in Liberman (1975) and Liberman and Prince (1977) and holds that, unlike other phonological properties, stress is not a feature; rather, it is the hierarchical rhythmic organisation of utterances. The central idea of Metrical Phonology (MP hereafter) is to capture the hierarchical nature of stress in a representation of its own, outside other segmental matrix that includes other features (Kager, 1995). The theory, applied to the phenomena of stress and syllabicity, exploits the notion of relations of constituency and relative strength or prominence between contiguous prosodic units (like syllables and stress groups). In other words, the basic assumption is that stress patterns reflect an underlying structure in which stronger and weaker constituents are juxtaposed (Clark et al, 2007:417). This relative prominence is expressed in the classical notation of MP using binary branching trees labeled S (strong) and W (weak) in which the two sisters of each branch are either [S W] (trochaic) or [W S] (iambic) with the more prominent syllable dominated by S and the less by W.
The labels – S and W – originate from a root ‘r’ and are relationally defined, thus: S means ‘stronger than W’ and W means ‘weaker than S’. Strong and weak syllables are paired by a procedure called foot formation and a node is strong only by virtue of the fact that it is the sister of a weak node. Binary structures representing phonological constituency in classical MP assume the following forms:
r r r
S W W S W S
mo ther en dure
S W S W
in to na tion
As in the generative approach, stress assignment in MP is rule-governed. There are two main rules:
1. Lexical Category Prominence Rule (LCPR) which operates on simple and compound words, and
2. Nuclear Stress Rule (NSR) which covers phrases and sentences.
Liberman and Prince (1977:271) state the LCPR as follows:
For any pair of sister nodes [N1, N2], If [N1 N2] L where L is a lexical category, then N2 is strong (S) if (iff) it branches.
The concept of branchingness is a metrical terminology referring to syllable quantity. A heavy syllable is said to have a branching rhyme and a light syllable, a non-branching rhyme. In compound words, branchingness refers to the numerical strength of the constituent nodes such that while N2 (master) in headmaster is strong as a consequence of having branched into two syllables (as against its sister – head), N1 (oil-tanker) in oil-tanker driver is strong since N2 (driver) does not branch into two new words. The application of the LCPR to simple and compound words is illustrated below:
r r r S
S W W S W S W
mo ther pre tend head ma ster
For the NSR, If [N1 N2]p where P is a phrasal category, then N2 is strong. Cruttenden (1986:31) explains the [N1 N2] p as the configuration [xy] in which y is strong. Y here refers to the rightmost elements in phrases and the predicates of sentences. The development of metrical theory has been on the basis of three distinct formalisms: one using metrical trees, another involving metrical grids and a third using ternary (SWS) structures. While Trees encode stress patterns, Grids reflect patterns of rhythm (Durand, 1990:234) and the ternary structures combine S-placement rules and rhythmic constraints to assign S and W directly to base forms and suffixes taking into account both syllable structure and morphological information.
We have, however, selected for our analysis in this study, a combination of Schane’s (1979) ternary notation and metrical trees. This is due to our conviction that these two complement each other – ternary structures capture the traditional rhythmic structuring of English utterances into feet (be it iambic or trochaic) thereby enhancing the strict relationality of early binary branching metrical tree notations. Schane had earlier argued that hierarchically ordered binary branching structures are ‘overly complex,’ while Durand (1990:235) also observes that binarism leads to a ‘proliferation of structures.’ Thus, ternary structures present ‘….a more abstract, yet simpler, characterisation of stress in terms of alternating strong and weak syllables’ (Schane, 1979:600).
Research Procedure
The study involved a total sample of fifty University of Lagos undergraduates, twenty-five of who are of Yoruba origin while the other twenty five are of Igbo origin. The sample was selected from the different faculties and levels using the Judgement/Purposive Sampling Technique. In addition to linguistic background, the variables of sex, age and socio-economic background were considered in selecting the sample. To collect the data for the study, a word list containing fifty polysyllabic words was presented to each respondent who read them into a SONY IC recorder. Each respondent was also engaged in a casual conversation designed to further identify the peculiarities of their stress realisation which might have been consciously concealed during the more formal recording process. These were surreptitiously recorded, played back and transcribed for statistical analysis.
The analysis involved counting the tokens of the variants of the items being tested, converting them into numbers and statistically calculating them to determine how often they occurred. These variant stress patterns were cross-checked against the Standard British English (SBE) pronunciation patterns as contained in Daniel Jones’ Cambridge English Pronouncing Dictionary (18th Edition) which was used as the control for appropriateness. The study demonstrates that the Igbo accent of NE has a stress assignment pattern that is markedly different from that of its Yoruba counterpart. The varying patterns were subsequently captured in the SWS and/or arboreal notations of Metrical Phonology where necessary.
The Sample
The sample of fifty students was selected from the entire University of Lagos (UNILAG hereafter) undergraduate student population. An equal sample was selected from each of the L1 Yoruba and L1 Igbo groups whose accents, in line with Igboanusi (2006), are referred to as Yoruba English (YE) and Igbo English (IE) respectively, in this study. Out of the fifty respondents, twenty-five (50%) were male while the other twenty-five (50%) were female. Twenty three (46%) were aged below 30, nineteen (38%) were between the ages of 30 and 49 while eight (16%) were aged 50 and above. Nineteen (38%) were from the Humanities, sixteen (32%) from the Sciences and fifteen (30%) from the Faculty of Education. While eleven (22 %) were in 100 level, twelve (24%) were in 200 level, another eleven (22%) were in 300 level while the remaining sixteen (32%) fell within 400 level and above. Table 3 below presents the demographic details of the fifty respondents:
Table 3: Demographic Characteristics of Respondents
| Variable | Igbo Respondents Frequency (Over 25) | Yoruba Respondents Frequency (Over 25) | |
| Sex | Male | 12(48%) | 13(52%) |
| Female | 13(52%) | 12(48%) | |
| Age | Below 30 | 11(44%) | 12(48%) |
| 30-49 | 09(36%) | 10(40%) | |
| 50 and above | 05(20%) | 03(12%) | |
| Faculties | Humanities | 09(36%) | 10(40%) |
| Sciences | 08(32%) | 08(32%) | |
| Education | 08(32%) | 07(28%) | |
| Level of Study | 100 | 06(24%) | 05(20%) |
| 200 | 06(24%) | 06(24%) | |
| 300 | 06(24%) | 05(20%) | |
| 400 and above | 07(28%) | 09(36%) |
The Data
Each of the fifty subjects was tested on a total of fifty items. The list consisted of potentially divergently-stressed English words with relatively uncertain NE stress patterns. Disyllabic words tested include PROtein, ANnexe, SCHEdule, CRAYon, BAPtist and SYrup. Trisyllabic words include comPOnent, umBRELla, CHAracter, seMEster, aGENda, SEminar, FAculty, SPIritual, HOSpital and CYlinder. Quadrisyllabic words tested include COMfortable, CEremony, VEgetable, TEStimony and DIFficulty while compounds include CLASSroom, TEXTbook, EARring and BOOKshop.
Statistical Analysis
The total number of test items for stress placement for each group was 25 x 50 = 1,250. Of the 1,250 items, Yoruba subjects were able to appropriately stress three hundred and eighty three (31%) while Igbo subjects were able to stress five hundred and fourteen (41%). Beyond the question of appropriateness, this study also focused on the differences in the stress preferences of the respondents. While IE demonstrated an unparalleled preference for initial stress, having stressed six hundred and eighteen items out of the one thousand, two hundred and fifty (49%) initially; YE’s preference was predominantly final stress – five hundred and forty three items out of one thousand, two hundred and fifty (43%). This result underlines the need for a reassessment of the status: ‘Nigerian English is treated as a homogeneous variety of English in terms of stress assignment…’ Atoye (2005). It demonstrates that IE and YE stress are individually different yet collectively different from RP. Table 4 below presents the subjects’ overall performance while Fig. 1 captures the individual accents’ stress preferences.
Table 4: Subjects’ Overall Performance in Word Stress Placement
| Group | Total No. of Items Tested | Initial Syllable Stress | Antepenultimate Syllable Stress | Penultimate Syllable Stress | Final Syllable Stress | Appropriately Stressed |
| Yoruba | 1,250 | 374 | 97 | 236 | 543 | 383 |
| Igbo | 1,250 | 618 | 53 | 133 | 446 | 514 |
Fig. 1: Stress Preferences of IE and YE Speakers
Analysis of Subjects’ Stress Placement on Word Types
The fifty test items comprised fifteen disyllabic words, fifteen trisyllabic words, ten quadrisyllabic words and ten compound words. The total number of test items for stress placement for each group was 25 x 50 = 1,250. The Yoruba subjects appropriately stressed one hundred and nine (29%) out of three hundred and seventy-five items of disyllabic words while Igbo subjects correctly stressed one hundred and forty one (38%) items. In the trisyllabic words category, Yoruba subjects correctly stressed one hundred and fifty one (40%) out of three hundred and seventy-five items while Igbo subjects stressed two hundred and eight items (55%) correctly. For quadrisyllabic words, Yoruba subjects appropriately stressed seventy (28%) while Igbo subjects stressed ninety nine (40%). For the compound words, Yoruba subjects appropriately stressed fifty three (21%) out of two hundred and fifty items while Igbo subjects scored sixty six (26%).The differentials suggest that IE stress patterning significantly varies with that of YE. While trisyllabic words are attested as the aspect with the highest degree of divergence, compound words show the highest level of convergence.Tables 5 and 6 below present Yoruba and Igbo subjects’ comparative stress placement of the different word types tested.
Table 5: Analysis of Yoruba Subjects’ Comparative Stress Placement of Word Types
| No. | Word type | No. of Items | Potential Score | Actual Score | Percentage |
| 1. | Disyllabic | 15 | 375 | 109 | 29 |
| 2. | Trisyllabic | 15 | 375 | 151 | 40 |
| 3. | Quadrisyllabic | 10 | 250 | 70 | 28 |
| 4. | Compound | 10 | 250 | 53 | 21 |
| Total | 50 | 1250 | 383 | 31 |
Table 6: Analysis of Igbo Subjects’ Comparative Stress Placement of Word Types
| No. | Word type | No. of Items | Potential Score | Actual Score | Percentage |
| 1. | Disyllabic | 15 | 375 | 141 | 38 |
| 2. | Trisyllabic | 15 | 375 | 208 | 56 |
| 3. | Quadrisyllabic | 10 | 250 | 99 | 40 |
| 4. | Compound | 10 | 250 | 66 | 26 |
| Total | 50 | 1250 | 514 | 41 |
Disyllabic Words
The analysis revealed that speakers of YE show a preference for ultimate syllable stress in disyllabic words as fourteen (PROtein, COLleague, SYrup, ANnexe, CRAyon, HIjack, PLANtain, SAlad, SCHEdule, inTACT, BAPtist, MATtress, VOmit and apPLAUSE) out of the fifteen items tested were stressed finally by them giving rise to the patterns: proTEIN, colLEAGUE, syRUP, anNEXE, craYON, hiJACK, planTAIN, saLAD, scheDULE, inTACT, bapTIST, matTRESS, voMIT and apPLAUSE. The attested IE stress pattern for disyllabic words demonstrated a consistent alternation between the final and penultimate syllables with penultimately-stressed items outweighing their ultimately-stressed counterparts. In all, six items (colLEAGUE, hiJACK, planTAIN, saLAD, matTRESS and voMIT) were stressed finally by IE subjects, while the rest were stressed on the initial syllable (SCHEdule, INtact, BAPtist, APplause, PROtein, SYrup, ANnexe, CRAyon and SUCcess). Convergent stress patterns were observed in a total of seven items (colLEAGUE, hiJACK, planTAIN, saLAD, matTRESS, voMIT and SUCcess) while eight (protein, syrup, annexe, applause, intact, schedule, Baptist and mushroom) were divergently stressed.
Trisyllabic Words
Fifteen trisyllabic words were tested. The analysis revealed that YE speakers have a preference for penultimate syllable stress in trisyllabic English words. Out of the fifteen trisyllabic words tested (comPOnent, umBRELla, CHAracter, seMEster, aGENda, uTENsil, CYlinder, FAculty, SPIritual, HOSpital, inTEStine, SIGnature, SEminar, INterview and ilLITerate), ten were stressed penultimately (comPOnent, umBRElla, chaRACter, seMESter, cyLINder, faCULty, spiRItual, hoSPItal, ilLIterate and sigNAture). Three attracted antepenultimate stress (Agenda, Utensil and INtestine while the remaining two attracted ultimate/final stress (semiNAR and interVIEW). IE speakers were more consistent in their trisyllabic word stress patterning. Their preference for antepenultimate syllable stress was evident as they stressed eleven out of the fifteen items of trisyllabic words antepenultimately giving rise to such patterns as COMponent, UMbrella, CHAracter, SEmester, CYlinder, FAculty, SPIritual, HOSpital, SIGnature, SEminar, and ILliterate. Three items attracted penultimate syllable stress – aGENda, uTENsil and inTEStine – while the remaining one – INterview – received ultimate stress (interVIEW). Interestingly, words which attracted penultimate stress in YE (comPOnent, umBRElla, chaRACter, seMESter, cyLINder, faCULty, spiRItual, hoSPItal and sigNAture) were consistently stressed antepenultimately in IE while those which received antepenultimate stress in YE (Agenda, Utensil and INtestine) were stressed penultimately in IE. Similar stress patterns were attested in interview (which was stressed finally) in both accents. In all, trisyllabic words demonstrated the highest level of divergence among the different word types tested.
Quadrisyllabic Words
Ten items of quadrisyllabic structure were tested – COMfortable, CEremony, VEgetable, TEStimony, conGRAtulate, indePENdent, imPOSsible, poLIticise, CRIticism and diVERsify. Our analysis of this category reveals that 50% of the words, which were stressed antepenultimately by YE respondents (comFORtable, ceREmony, veGEtable, teSTImony and criTIcism), were stressed initially by IE respondents (COMfortable, CEremony, VEgetable, TEStimony and CRIticism). Shared patterns attested include congratuLATE, INdependent, IMpossible, politiCIZE and diversiFY.
Compound Words
The highest level of convergence was observed in this category. Out of the ten items tested (TAble mat, MATCHmaking, PILlowcase, CLASSroom, BEDspread, TEXTbook, PADlock, WEDding ring, BOOKshop and EARring), six were uniformly stressed on the final constituent by both groups (table MAT, match MAking, pillow CASE, bed SPRead, padLOCK and wedding RING). Divergent patterns were observed in classroom, textbook, bookshop and earring each of which was stressed ultimately by YE respondents (as classROOM, textBOOK, bookSHOP and earRING) and penultimately by IE respondents (CLASSroom, TEXTbook, BOOKshop and EARring). The combined results of the Yoruba and Igbo subjects’ realisation of stress in English compounds corroborate Kujore (1985) and Jowitt’s (1991) findings that in articulating English compounds, NE has a tendency to shift stress as far to the right as possible.
Metrical Analysis
In this section, the dominant stress patterns observed in each of the accents studied are described in metrical terms. Two major contrasts were observed in the disyllabic words category: S W (trochaic) and W S (iambic) structures. The SW patterns were predominantly favoured by the IE respondents while YE respondents favoured WS patterns. An interesting dimension to the disparity observed between the two accents is the unconscious attempt to strengthen the vowel in the stressed syllable thus imposing that characteristic branching quality of stressed syllables on it. A few examples will suffice here: the word protein is realized as /prəʊtin/ in RP; the diphthong /əʊ/ assigning its branchingness, and stress, to the first syllable (SW). In YE, the same word is realized as [proten] – [e] being a monophthongized version of the diphthong /eɪ/, hence, a tense vowel. This substitution of [e] for /i/ in the ultimate syllable of ‘protein’ fulfills the conditions necessary for its realization as W S – LCPR. Similar patterns were also observed in ‘syrup’ and ‘crayon’ where the tense [ɔ:] replaced /ʌ/ and /ə/ respectively in the final syllables thus giving rise to W S as against S W structures.
In the trisyllabic word category, two distinct metrical patterns were also observed – WSW (ambibranch) patterns which were preferred by YE speakers and SWW (dactylic) patterns preferred by IE speakers. These patterns indirectly retain the trochaic and iambic patterns preferred by the IE and YE respondents respectively as discussed above, when analysed in line with the binary branching metrical structures as shown below:
IE YE
r r
S S
S W W W S W
cha rac ter cha rac ter
The metrical trees show that in both accents, the first two nodes that emerge from the root are in
conformity with the identified preferred patterns. The subsequent divisions simply enhance the prominence of the Designated Terminal Element (here ‘cha-’ in IE but ‘rac-‘ in YE) while subduing the sister node. Similar patterns are upheld in comPOnent/COMponent, umBRELla/UMbrella, seMEster/SEmester, cyLINder/CYlinder, faCULty/FAculty, spiRItual/SPIritual, hosPItal/HOSpital and sigNAture/SIGnature. Only one anapestic pattern (WWS) was observed in YE in the word seminar – semiNAR – which was realized dactylically (SWW) in IE as SEminar. A close analysis of the quadrisyllabic items reveals similar trochaic patterns at the root of the tree, also similar to what obtains in the target accent. However, as these original nodes further branch into stronger and weaker nodes, the individual accents’ preferences noted above resurface, specifically within the first two syllables as demonstrated below:
RP IE YE
r r r
S W S W S W
S W S W S W S W W S S W
com fort a ble com fort a ble com fort a ble
The structures above are sustained in VEgetable/ veGEtable, TEstimony/ teSTImony and CEremony/ ceREmony. Looking at the word ‘comfortable’, one would observe that the structures have implications for the stress quality of the affix ‘–able’ in each of the accents. Expectedly, the metrical patterns of the divergently stressed compound nouns reflect the established preferences of the accents. While YE has WS structures in classroom, textbook, bookshop and earring, IE has SW, leading to the contrasts: classROOM/ CLASSroom, textBOOK/ TEXTbook, bookSHOP/BOOKshop and earRING/EARring.
Discussion of Findings
The findings of this study clearly confirm Roach’s (2000:102) assertion that English word stress and the phonological rules which derive them are “simple enough in theory but highly complex in practice.” Despite the linguistic affinity between Igbo and Yoruba and, by implication, IE and YE (Jowitt, 1991:71) – both Southern accents of Nigerian English – the stress systems, based on the analysis carried out in this study, have revealed a high level of variation which defies etymological analysis. One interesting aspect of the disyllabic words category is the convergence in the accents’ realization of the word sucCESS as SUCcess. This is apparently in line with the Noun-Verb Alternation (NVA) stress rule in NE and Cameroon English (CamE) (Simo-Bobda, 2010) which states that ‘words which have a nominal form and a verbal form are stressed initially in their nominal form and finally in their verbal form.’ Unfortunately, while this rule stands for such pairs as IMport (N)/ imPORT (V) and PROduce (N)/ proDUCE (V); sucCESS (N)/sucCEED (V) alongside exTENT (N)/ exTEND (V) and preTENCE (N)/ preTEND (V) constitute exceptions. Hence, there is a vacuous application of this rule in both accents for the word success. However the fact that apPLAUSE was divergently stressed, the possible alternation of its stress pattern with that of its verbal form apPLAUD notwithstanding, makes it impossible for us to make any generalisation on the validity of the NVA in the two accents studied.
Furthermore, given the differences between the stress systems of these accents, on the one hand, and between the RP stress system and those of these accents, on the other, neither IE stress nor YE stress can be said to totally replicate RP stress. This is a pointer to the domestication of English stress. However, the disparity between the two accents calls for concern particularly when viewed in the light of its implication for the emergence of the Standard Nigerian English Accent. Crystal (2008:450) views a standard language/dialect/variety as that variety which ‘cuts across regional differences, providing a unified means of communication….’ The differences in the respondents’ choices of stressed syllables in the disyllabic (e.g. PROtein ~ proTEIN, BAPtist ~ bapTIST), trisyllabic (e.g. SEmester ~ seMESter, UMbrella ~ umBRElla), quadrisyllabic (e.g. COMfortable ~ comFORtable, VEgetable ~ veGEtable) and compound words (e.g. TEXTbook ~ textBOOK, EARring ~ earRING) tested present us with another task of deciding between the variant stress patterns of each word, which to adopt in the NE corpus.
Conclusion
This study has demonstrated the need for the NE scholar to expand the frontiers of his analysis of the variety by complementing the existing studies on its segmentals with empirical studies on the suprasegmental details (which are equally essential to the description of any accent). It is only the result of such studies that can guarantee the reliability or otherwise of the generalisations we often make about the NEA particularly with regard to the prosodic feature of stress. As a starting point, it is recommended in this study that Nigerian English scholars pay more attention to the stress patterns of (at least) the three major accents of NE. This will enable us to identify the convergences which will truly be representative of the NEA while still retaining the divergences as individual accent markers. Forming our opinions of this variety of English based on localised studies will, at best, be deceptive. Thus, this study has dire implications for the standardisation of Nigerian English pronunciation. The indices suggest that this may not necessarily involve a natural development but an imposition of particular patterns as the standard (the educational parameter which has often been used to determine the standard having failed, as far as pronunciation is concerned). Since, naturally, such an imposition cannot be devoid of codification, this study serves as a wake-up call to all NE scholars to the task ahead. Ensuring the success of the description, codification and, by implication, standardisation of the NEA is a task for all.
References
Adetugbo, A. (2004): “Problems of Standardisation and Nigerian English Phonology” in Dadzie, A.B.K. and Awonusi, S. (eds.) Nigerian English Influences and Chracteristics. Lagos: Concept pp 179-199.
Akinjobi, A. (2004): “A Phonological Investigation of Vowel Weakening and Unstressed Syllable Obscuration in Educated Yoruba English”. Unpublished Ph.D Dissertation, University of Ibadan, Nigeria.
Anyagwa, C. (2013): “Word Stress in Nigerian (Igbo) English”. Unpublished Ph.D. Dissertation, University of Lagos, Lagos, Nigeria.
— (2015): “Word Stress in Nigerian English: A Comparative Study”, in Opeibi et al.(eds.) Essays on Language in Societal Transformation: A Festschrift in Honour of Prof. Segun Awonusi. Research in English and Applied Linguistics (REAL STUDIES) Vol 9 pp. 41-54. Gӧttingen: Cuvillier.
Atoye, R. (1991): “Word Stress in Nigerian English.” World Englishes 10:1–6.
— (2005): “Word Stress in Nigerian English and the Challenges of Nigerian English Oral Literacy.” In Proceedings of the 22nd Annual Conference of the NESA pp 39-46.
Awonusi, S. (1987): “Regional Accents and Inherent Variability in Nigerian English: A Historical Analysis” English Studies, 67.6: 555-560.
— (2004): “Some Characteristics of Nigerian English Phonology.” In Nigerian English Influences and Chracteristics. Eds. Dadzie, A.B.K. and Awonusi, Segun. Lagos: Concept,. pp 203-225.
Banjo, A. (1979). “Beyond Intelligibility: A Presidential Address”. In Ubahakwe, E. (ed.) Varieties and Functions of English in Nigeria pp. 7-13. Ibadan: African University press.
Bolinger, D. (1981): “Two Kinds of Vowels, two Kinds of Rhythm” in Indiana University Linguistic Club.
Clark, J. et al, (2007): An Introduction to Phonetics and Phonology. 3rd edn. Oxford: Basil Blackwell.
Cruttenden, A. (1986): Intonation. Cambridge: CUP.
Crystal, D. (2008): A Dictionary of Linguistics and Phonetics. Oxford: Blackwell.
Dadzie, A.B.K. (2004): “The Concept of Nigerian English” in Dadzie and Awonusi (eds.) pp 85-99.
Durand, J. (1990): Generative and Non-Linear Phonology. London: Longman.
Gut, U. (2001) Gut, Ulrike: “Prosodic Aspects of Standard Nigerian English”. http://wwwspectrum.uni-bielefeld.de.TAPS/Gut.pdfWeb. Accessed 9/12/2009.
Gut, U. and Milde, J. (2002): “The Prosody of Nigerian English”. In ISCA Archive: Speech Prosody. http://www.isca-speech.org/archive Web. 30/6/2010
Igboanusi, H. (2006):“A Comparative Study of the Pronunciation Features of Igbo English and Yoruba English Speakers of Nigeria” English Studies, 87.4: 490 – 497.
Jibril, M. (1979): “Regional Variation in Nigerian Spoken English.” In Varieties and Functions of English in Nigeria. Ubahakwe, Ebo. (Ed.) Ibadan: A.U.P./NESA, pp 43-53.
— (1982): “Phonological Variation in Nigerian English”. Unpublished Ph. D. Dissertation,University of Lancaster.
Jolayemi, D. (2008): “Acoustic Analysis of Tone and Stress: An Illustration with Varieties of English Spoken in Nigeria.” In Focus on English: Linguistic structure, Language variation and Discursive Use. Wolf, H. et al (Eds.). Leipzig: Leipzig University Press pp 105-119.
Jowitt, D. (1991): Nigerian English Usage: An Introduction. Zaria: Longman.
— (2000): “Patterns of Nigerian English Intonation.” English World-Wide 21: 63–80.
Kager, R. (1995): “The Metrical Theory of Word Stress.” In A Handbook of Phonological Theories. Goldsmith, J. (Ed.) Oxford: Blackwell, 1995:367-402.
Kujore, O. (1985): English Usage: Some Notable Nigerian Variations. Ibadan: Evans Brothers.
Liberman, M. (1979): The Intonational System of English. MIT Ph.D Dissertation. New York: Garland.
Liberman, M. and Prince, A. (1977): “On Stress and Linguistic Rhythm.” Linguistic Inquiry 8:249-336.
Odumuh, A. (1987): Nigerian English. Zaria: ABU Press
Olaniyi, O.K. (2013): ‘The Taxonomy of Nigerian Varieties of Spoken English”. www.academicjournals.org/article/article1411984705_Olaniyi.pdf Accessed 1/10/14.
Omachonu, G. (2008). “Comparative Optimality Theory Analysis of Primary Stress Assignment in Standard British and Nigerian English.” Journal of Language & Translation 9.1: 91-112.
Osifeso, A. E. (2017): “Word Stress among Yoruba L2 Speakers of English in Lagos: An Optimality Approach.” MAJELS 2 (10), 19-28
Osuafor, C. (2002): The English Language in Nigeria. Owerri: Great Versatile Publishers.
Roach, P. (2000): English Phonetics and Phonology. Cambridge: CUP.
Schane, S (1979) “The Rhythmic Nature of English Word Accentuation.” Language 55.3: 559–602.
Simo Bobda, A. (2007): “Some Segmental Rules of Nigerian English Phonology.” English World-Wide 28.3: 279–310(b).
— (2010): “Word Stress in Cameroon and Nigerian Englishes.” World Englishes 29.1:59–74.
Tiffen, B. (1974): “The Intelligibility of Nigerian English”. Unpublished Ph.D Dissertation, University of London.
Udofot, I. (2003): “Stress and Rhythm in the Nigerian Acccent of English: A Preliminary Investigation.” English World-Wide 24.2 pp 201-220.
Wells, J. (1982). Accents of English vols. 1-3. London: Arnold.