Ethnolinguistic Heterogeneity in Cameroon English Pronunciation

Abstract

The investigation of ethnolects or ethnic varieties of languages has been around for a while now. While most previous studies have focused on immigrant ethnolects, the present work focuses on an indigenised variety of English, Cameroon English, and how the phonological (pronunciation) features of indigenous Cameroonian languages are represented in English. This study, therefore, lists and describes the major features of two ethnolects of Cameroon English: Nso’ English and Wimbum English. Using data collected from primary school children and university students and graduates, the chapter identifies and describes processes such as diphthong reduction or simplification, vowel lowering, and vowel shortening. Though these processes are common in most ethnolects, their realisation and the vowel phonemes affected are different and specific to each ethnolect.

Given that these vocalic processes resemble processes in the indigenous languages, a possible reason for their persistence in these ethnolects, this chapter illustrates, is substratum influence. Because the major ethnolectal features are also used by acrolectal speakers (i.e. university students and graduates), they are considered here authentic markers of these ethnolinguistic varieties, and pointers to the heterogeneity in (ethnic) accents in English in Cameroon.

1 Introduction

Previous studies on ethnic varieties or ethnolects of English in Cameroon have identified both similar and dissimilar ethnolinguistic features , especially in pronunciation. The aim of this chapter is to describe those salient phonetic features which identify ethnic-influenced pronunciation within Cameroon English (CamE) , and to account for their retention in the speech of educated, acrolectal speakers . The dialectological approach is used to describe the variation in the pronunciation of English segmental features by some ethnic group members in Cameroon. Substrate influence from the ethnic languages is the major source of the variation witnessed, and is used here to validate the presence of ethnolects in CamE pronunciation (see also Fonyuy 2012) . The ethnolectal speech is described in relation to CamE and RP pronunciation norms, CamE being the input model and RP the de facto reference for teaching English in Cameroonian schools .

Using empirical evidence from data collected in Cameroon in 2010, this chapter illustrates firstly, that ethnic heterogeneity in CamE pronunciation is verifiable in the speech of speakers who double as L1 speakers of a Cameroonian ethnic or indigenous language and as L2 speakers of CamE . Secondly, the level of education and exposure to Standard Englishes impact the ethnolectal pronunciation choices. That is, even though the vocalic processes focused on in the case studies, Nso’¹ and Wimbum² ethnic Englishes, are systematic, they vary with speakers’ level of education. In spite of the educational impact, some of these ethnic language features are retained in acrolectal speech . This type of retention, occurring in spite of level of education and exposure, signals variety or ethnolectal stabilisation, and also captures the degree of ethnolinguistic heterogeneity in CamE pronunciation.

The linguistic plurality of Cameroon makes it an interesting destination for various types of linguistic investigations. The country has a population of about 20 million (National Population Census 2005) belonging to around 200 ethnic groups, who speak over 277 indigenous languages (Lewis 2009) . There are two exogenous languages , English and French, which function as co-official languages of education, media, law, administration, etc. Additionally, there is the widespread Cameroon Pidgin English (CPE) serving as a language for wider, inter-ethnic communication, and the bilingual mixed language, Camfranglais , spoken mostly by the youth. Given this multilingual background, bilingualism or multilingualism is the norm for most Cameroonians, making heterogeneity in pronunciation almost inevitable. CamE pronunciation , therefore, exhibits some ethnolinguistic features which could be attributed to this multilingual status of the speakers. General references are made in this chapter to the Nso’, Kom, Bafut, Bakossi, Moghamo, and Wimbum ethnic communities and the common linguistic variables identified with the way they speak English, while the case study focuses closely on two ethnolects : Nso’ English (NsoE) and Wimbum English (WimE) .

2 Ethnolects and Substrate Influence

Although the place of substrate influence has been disputed in immigrant ethnolects , it seems inevitable in non-immigrant multilingual communities where languages play different roles in people’s daily lives—as official languages, home languages, etc. Though inherently controversial in certain contexts, in most postcolonial multilingual communities , “differences in linguistic behaviour among ethnic groups are usually assumed to result from substrate transfer from the languages originally or still spoken by each ethnic group” (Hoffman and Walker 2010, p. 42) . In some cases, these ethnic variations give the local variety of English its new ecological distinctiveness.

Ethnolinguistic diversity in mainstream national varieties of English also exists in other postcolonial communities other than Cameroon. For instance, Adjaye (2005, p. i) presents ethnolinguistic heterogeneity in Ghanaian English , stating that although “Akan languages have phonemic distinctiveness of /u/ and /ʊ/ as well as /i/ and /ɪ/ … only the latter is maintained in Akan speakers’ English.” This suggests that ‘ship’ and ‘sheep’ are synonymous in the English of Akan speakers. In Nigerian English , Banjo (1996, p. 76) gives the phonological bases for ethnolinguistic variation in Nigerian English noting that

[n]one of the three major Nigerian languages (Hausa, Igbo and Yoruba) has up to ten pure vowels, and only Hausa has a diphthong…and in many cases, RP diphthongs are monophthongised…While some languages (like Hausa and Igbo) do have the phoneme /z/, others like Yoruba do not have it and so a Yoruba Variety 1 [basilectal] speaker would substitute /s/. Similarly, for the speakers of some languages (again including Yoruba) /f/ is substituted for /v/ and /ʃ/ for /ʧ/.

In correlating ethnic pronunciation patterns to sociolectal hierarchy, Banjo (1996, p. 76) explains that some of the features, especially, homophony due to the absence of vowel length, occur more in basilectal speech than in acrolectal speech . For instance, according to him, in Nigerian English , “Variety 1 [basilectal] /li:v/ is equivalent to RP /li:v/ and /liv/ (leave and live)…, /faja/ represents RP /faiə/ (fire).” However, he (Banjo 1996, p. 79) further states that

Variety III…represents the acrolectal use of English in Nigeria…As noted elsewhere (Banjo 1971) , it shares the same deep structure with RP but has Nigerian phonetic features. Sometimes these phonetic features, though not impeding intelligibility , are strong enough to mark the speaker’s provenance.

Using data from Tswana English , Van Rooy (2002, p. 148, 154) explains that in South African English, there is the /a/→/α/ variation in medial positions in the words /arrant/ (around), /abαt/ (about), and /dinαnsas/ (denounces). It is the impact of Tswana phonology, Van Rooy suggests, that models this variation.

These ethnic variations indicate how important substrate influence is to the emergence of not only ethnolects but also mainstream national varieties. Interestingly, the segmental features affected are similar across ethnic languages and varieties of English. These include diphthong simplification , the absence of vowel length distinctions, and the substitution of voiced sounds for voiceless sounds.

From all these illustrations, it is evident that ethnic heterogeneity in African Englishes is based on substrate influences , which may or may not disappear along the speakers’ educational achievement or exposure to other varieties of English. The following section locates ethnolinguistic heterogeneity in CamE pronunciation by comparing ethnolectal features to CamE and RP. It also reviews some related literature on CamE ethnolects .

3 Ethnolinguistic Heterogeneity in CamE Pronunciation: An Overview

Although the phonologies of some Cameroonian indigenous languages form part of the base of CamE phonology , some of their consonants are significantly different from CamE consonants, especially coarticulated consonants as in Lamnso’ <dz>, <kp>, <gb>, <dr> and Limbum <rk>, <rts>, and <rb>. The realisation of RP <and> and CamE <an> as <andr> in the NsoE accent is a result of the absence of the phoneme /d/ in word-final positions in Lamnso’. Lamnso’ speakers of English, therefore, introduce this structure into English since it is identical to another structure in their L1 phonology. Despite such differences, CamE shares a number of similar vowel features with most of its ethnolects , but as summarised in Table 6.8, vowels in the same word environment undergo different types of phonetic modification in some of the ethnolects . The phonemes of CamE and some ethnolects are also significantly different from RP especially in vowel length, voice quality, and the permissibility and realisation of diphthongs and triphthongs. CamE and RP generally share the same consonant features, except for the RP inter-dental fricatives [θ, ð] where RP [θ]in (thin) and [ð]em (them) are produced in CamE respectively as [t]in and [d]em (see Simo Bobda and Mbangwana 2004, pp. 200–204) .

Different researchers have analysed deviations from RP in ethnic English pronunciation in Cameroon. For instance, the variant forms of pronunciation in the English speech of some educated Nso’ people are discussed by Yusimbom (1992) . Under the sub-title use of foreign vowels, Yusimbom (1992, p. 47) states that “[i], [a], [o] are vowels foreign to English” and cites the example in which educated Nso’ speakers “use [wil] for English/wɪl/ <will>.” Other pronunciation phenomena discussed include vowel insertion and monophthongisation. In analysing the possible causes of Lamnso’ influence on NsoE and the negative effects that such interference has on intelligibility , Yuyun (1996) revisits some aspects of NsoE identified by Yusimbom (1992) . In addition, Yuyun (1996, p. 44) arrives at the finding that “words with /i:/ preceded by the nasals /m/ and /n/ are rare in Lamnso’, so that NsoE bilinguals replace English /i:/ by /e/”³ yielding realisations like the following:

Word	Nso English	RP
meat	/met/	/mi:t/
neat	/net/	/ni:t/
knee	/ne/	/ni:/

Similarly, Sala (1999) investigates some major vowel alternation processes in Lamnso’ speakers’ English. Using vertical and horizontal counts, Sala (1999, p. 36) identifies different environments in which the /i/ for /e/ alternation phenomenon occurs. He explains that it is as a result of “alternation and neutralisation of [i] and [e] after nasals within Lamnso’ itself.” Focusing on the evolution of some vowel pronunciation features in NsoE associated with level of education, Fonyuy (2003, 2013) proves that along the lectal continuum, i.e. from the basilect up to the acrolect, ethnic English features become less recurrent, but do not disappear completely even when speakers advance in education. From these contributions, the substrate effect of Lamnso’ phonology on English seems significant.

Not much research has been done on the English accent of the Wimbum. Nforgwei (2004, p. 12) briefly mentions the sociolinguistic aspect of the English spoken by the Wimbum: “Today, we can hear native speakers say ‘sútɛ’, pidgin word for ‘until’, ‘háràbàda’ fulfulde word for ‘until’ and ‘ìntí’ for the English word ‘until’.” This is an interesting case because we identify the WimE [i, u, ɛ] features not only in English, but also in CPE. Although the Nso’ and Kom ethnic groups also realise the word as ‘suté’, CPE generally realises it as ‘soté’ with the main variation being [u] → [o]. Neither CamE nor the NsoE speakers realise ‘ìntí’, which is characteristically a WimE feature. However, a more extensive study is that of Tamfu (1989) , who analyses vocalic variation in the spoken English of some educated Wimbum people. With the exception of WimE [ɛ] and [u] for RP /eɪ/ and /əʊ/ respectively, most of the variables he analyses are similar to CamE. On diphthong reduction, his example of t[ɛ]ble for t[eɪ]ble is an exclusive feature of the WimE accent (see Tamfu 1989, p. 109 ff.) . CamE rather realises it as t[e]ble.

Even less has been done on the Kom English accent. One reason may be that it has over the years evolved towards mainstream CamE pronunciation . But from folk discussions, the Kom are remembered for substituting /l/ for /r/ or inserting an additional vowel /i/ after an existing, mostly, back vowel between consonants in English. This creates diphthongs like [ai] and [oi]. One noted example of this process is Kom English realisation of l[ai]ndl[oi]rd for RP l[æ]ndl[ɔ:]d, and CamE l[a]ndl[o]rd (landlord). An example of this folk reference to Kom English is the following quotation from a Cameroonian online community. The writer is from Kom and so tries to use some of the features identified with the Kom, more for humour on the forum:

I am coming from Bikom land … The journey was not easy because … lain loba nobi fit loin for sain sain.⁴ (English: land rover could not run on sand)

Two main features of Kom English can be identified in this excerpt. First, there is the substitution of /l/ for /r/ noticeable in [r]over → [l]o[b]a and run → [l]oin. Second, there is an epenthetic /i/ inserted after the vowels in s[æ]nd → s[ai]n and r[ʌ]n → l[oi]n, resulting in the diphthongs [ai] and [oi].

Song (1996) investigates phonological processes such as vowel substitution and vowel insertion in the spoken English of some educated Kom people. The vowel insertion feature which is often quoted as a marker of the Kom English accent is absent in Song’s work. She (Song 1996) illustrates that vowel insertion is rare in the spoken English of most educated Kom people. From this, it can be hypothesised that the Kom English features which folks often refer to is a phenomenon of non-educated Kom speakers . In any case, it still remains a marker of ethnolinguistic variation in CamE pronunciation .

On his part, Masanga (1983) analyses the spoken English of educated Moghamo people with focus on vocalic processes such as vowel insertion, substitution and the introduction of foreign vowels. His results show that Moghamo English introduces an epenthetic vowel in an environment where CamE generally does not. An example he advances is Moghamo English quic[ki]ly for RP and CamE quic[k]ly (see Table 6.8).

In spite of all these attempts at describing ethnic English accents in Cameroon, not much has been done on these ethnic variations as belonging to, or departing from, mainstream CamE pronunciation . This gap is what this chapter endeavours to fill. Apart from the generative perspective on CamE by Simo Bobda (1994) and Simo Bobda and Chumbow (1999) , who use the “trilateral process” to analyse CamE phonology, most researchers on CamE and its ethnolects attribute these variations in CamE pronunciation to substrate influences . The analysis below follows a similar substrate approach with the aim of illustrating the heterogeneity that exists.

Noteworthy is that Hoffman and Walker (2010) dispute the substrate influence explanation basing their argument on generation gap and the absence of empirically tested interpretations. Their argument is substantial in that, firstly, they use a multi-generational approach in analysing ethnolects in immigrant communities in Canada, where English or French and not ethnic languages dominate daily speech, and immigrants are likely to take on the speech patterns of their host country. It is obvious that in such a context, the English speech of first generation speakers will exhibit more ethnic features than that of subsequent generations, in whose English, ethnic features may have disappeared. Secondly, their criteria for eligibility of informants are based more on ethnic decent and ethnic affiliation and not so much on an ethnic language as L1.

So, while Hoffman and Walker’s (2010) perspective on substrate influence is valid for the Canadian context, the Cameroonian is different. The respondents are not immigrants and are hence not pressured to integrate linguistically or otherwise; they live in communities where their native language is often dominant especially in informal, home domains; and there is no generational gap in the transmission of the native language. English is used in education and official and formal employment domains. I return to this briefly again in Sect. 6.4.3 below. The next section presents a brief overview of the ethnolinguistic history of Lamnso’ and Limbum.

4 Ethnolinguistic Overview of Lamnso’ and Limbum

Like a number of other ethnic groups in the Bamenda Grassfields, the Nso’ and the Wimbum trace their origin to the Tikar area, which lies to the North–East of the Bamun territory in the North–West Region of Cameroon. The main differences between Limbum and Lamnso’ languages are, while Limbum belongs to the Mbam Nkam, Nkambe cluster of languages, Lamnso’ belongs to the Ring, East cluster. Unlike Limbum which has more dialects, Lamnso’ is more homogeneous.

4.1 Limbum

Limbum is classified as a Niger-Congo, Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Southern, Eastern Grassfields Bantu, Mbam-Nkam, Nkambe (see Fransen 1995; Fowler and Zeitlyn 1996; Nforgwei 2004) . It has several dialects, though linguists do not agree on the exact number. Fiore (1987) identifies three dialects while Nforgwei (2004, p. 10) identifies four, i.e. Linti, Liwarr, Liyaa, and Lintumbaw. The main differences between these dialects are consonantal. However, the major ethnolectal features in WimE are rather at the level of vowels, not consonants. This means, therefore, that dialectal variation in Limbum itself does not affect the classification or types of ethnolectal features that surface in the speakers’ production of English.

4.2 Lamnso’

Lamnso’ is a Benue-Congo language that belongs to the Ring group of the Western Grassland Bantu group (see Grebe and Grebe 1976) . A more extended classification describes it as a Niger-Congo, Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Southern, Wide Grassfields, Narrow Grassfields, Ring, East.

Lamnso’ is not as heterogeneous as Limbum. For instance, Grebe and Grebe (1976) observe that Lamnso’ does not have any dialectal variation. So, the features identified with its speakers of English apply to all geographical regions where the language is spoken. An interesting aspect of its phonology is that, Lamnso’ modifies the vowels in borrowed words to suit its phonology, especially the realisation of /əʊ/ as [u]. Examples of some English loans that undergo this modification are /windu/ (window) and /tumatus/ (tomatoes).

4.3 The Uses of Endogenous and Exogenous Languages in Mbum and Nso’

As mentioned earlier, English and French are the official languages of Cameroon, and are hence used as medium of education, and for other official, formal business in administration, media, law, and formal employment. Speakers of Lamnso’ and Limbum are, therefore, exposed to these languages, but more extensively to English since it is used more in the anglophone part of the country. In these two locations, French is not regularly used outside the classroom as compared to English, and by extension CPE. There is also an extensive scenario of language contact in Mbum and Nso’ involving, besides these two languages, English, CPE, Fulfulde, French, and for parts of Mbum, Lamnso’. This is due to cosmopolitanism in Nkambe, a major socio-political town in Donga-Mantung Division and the vibrant economy of Ndu, a smaller town in the division. In addition to these, many Wimbum people also speak Fulfulde, spread mostly through Islam. Interestingly, Lamnso’ is also spoken in the South of Mbum which shares geographical boundary with Nso’. A linguistic outcome of this could be the similarity between the two groups in the pronunciation of some English phonemes.

The Nso’ people, on the other hand, make a unique difference in language loyalty. They are loyal to Lamnso’, their ethnic L1, since they use it very often among themselves in almost all domains of society, e.g. at home, in the market, on the farms, at play, in church, for business, for broadcast on community radio stations, and in workplaces. Here, English and CPE are used mostly for communication with non-natives. This loyalty to, and constant use of, Lamnso’ could explain why certain ethnolectal features persist in NsoE irrespective of exposure to other varieties and education.

Substrate influence , being one of the rudimentary reasons for the existence of ethnolects or ethnolinguistic variation , is central here. The contact of languages, especially between Limbum and Lamnso’, is also important because it sets the pace for variation or similarity in the patterns the two groups produce English sounds. Before illustrating the phonological phenomena that these groups make use of in English speech, let me present the data collection method I used to elicit the data.

5 Data and Methodology

The data in the case study of NsoE and WimE accents used in this chapter are a portion of the data I collected in Cameroon in 2010 for my PhD research project. As already indicated in Sect. 6.4, both ethnic groups are located in the North West Region of Cameroon. As the description of the informants in Table 6.1 shows, I used primary school children, considered loosely here as representing the basilectal level, and university students and graduates, representing acrolectal speech . The aim was to check if ethnolectal features disappear as speakers acquire more education and are exposed to other varieties of the language.

Table 6.1 Description of informants

In all, 79 informants were involved, with almost the same number of males and females.⁵ The informants were from Nso’ (39) and Mbum (40). They also had either Lamnso’ or Limbum as their first language.

The data were collected through recorded interviews comprising a series of casual but tactfully structured questions, which demanded answers at sentence level. In each carrier sentence, a discrete word containing a target phoneme was embedded. These phonemes are variables or observable ethnic features which mark out the ethnic variation in NsoE and WimE pronunciation. Tables 6.2 and 6.3 show the interview questions used for the NsoE and WimE respondents.

The sounds targeted in the NsoE speakers were the diphthong /ǝʊ/ and the long front vowel /i:/ in the words ‘spoke’ and ‘meal’ (Table 6.2). In the case of the WimE speakers, focus was on the diphthong /eɪ/ and the short high front vowel /ɪ/(Table 6.3).

Table 6.2 Interview questions for the NsoE respondents

Table 6.3 Interview questions for the WimE respondents

The interviews were recorded using a digital micro track II recorder, and later transcribed and quantified using wave and MSXL programmes. The descriptive statistics method was used to find the frequency of the phonetic variants realised by the different respondents. In the analysis, attention is paid to both educational level and ethnic origin, as possible reasons for variation in pronunciation on ethnolinguistic lines.

6 Ethnolectal Features in NsoE and WimE: A Case Study

This section identifies some of the major features linked to these two ethnic varieties with the help of the data collected using the interviews. Three vocalic processes are taken up here: diphthong reduction, vowel shortening , and vowel lowering . In some cases, two processes occur together, for instance, in WimE , the diphthong /eɪ/ is first of all reduced to a monophthong and then lowered to /ɛ/. The results in Tables 6.4, 6.5, 6.6, and 6.7 are used as empirical evidence for the existence of ethnolinguistic variation in CamE pronunciation . Although only a few processes are illustrated here, ethnic-specific linguistic processes are also noticeable in these and other ethnolects at the level of vowel raising and lowering, vowel alternation , phonemic coalescence, epenthetic vowel insertion, and disyllabification of triphthongs. For more on these and other processes in African Englishes, see Jowitt (1991) , Arua (1999) , Van Rooy (2002) , and Simo Bobda (2007) .

6.1 Diphthong Reduction

Diphthong reduction or simplification is a common phonological feature of most New Englishes . From Asia across to Africa, it has been attested in various varieties by Platt et al (1984) , Bamgbose et al. (1995) , and by Simo Bobda (1994) in the case of CamE . For Simo Bobda (2007, p. 412) , “All English diphthongs are prone to monophthongisation in African Englishes .” Reduction occurs in different ways in different varieties of English. For example, a diphthong could either be reduced to one of its phonemes, e.g. /eɪ/ to either [e] or [i], or to a completely different phoneme, as in /eɪ/ to [ɛ] in WimE discussed below.

Diphthong reduction is a shared pronunciation pattern of CamE and its ethnolects , but the segment which substitutes the diphthong and the word environment in which this occurs are specific to each ethnolect. As shown in Tables 6.4 and 6.5, NsoE and WimE reduce diphthongs to different sounds. The diphthong tested among the NsoE respondents is /əʊ/ (Table 6.4).

Table 6.4 Nso English: spoke → sp[u]ke (n = 39)

From Table 6.4, it is seen that 60 % of the 39 respondents of the NsoE sample reduce RP /əʊ/ in sp[əʊ]ke not to CamE /o/, sp[o]ke (32 %) but to NsoE [u], sp[u]ke. A possible reason for this is the absence of the diphthong /əʊ/ in Lamnso’. If one follows the speech learning model (see Flege et al. 1997) , which posits that if a phoneme is absent from the phonology of a parent language, then it will be difficult to perceive and realise it in the second language, then it is obvious that the reduction of /əʊ/ to [u] in NsoE is a substrate influence . Although /o/ and /u/ both exist in Lamnso’, there seems to be no phonetic contrast between the two sounds in the phonology of NsoE where they both merged into [u].

Also worth noting from the results in Table 6.4 is the high percentage for CamE realisations (32.5 %) and RP (5 %), especially among the acrolectal (tertiary level) respondents. These two percentages suggest that education and exposure certainly have an impact on the evolution of ethnolectal features. The basilectal speakers (primary level) have a much higher occurrence of ethnolectal features (Tables 6.4 and 6.5) than the acrolectal speakers . In Table 6.4, the reduction of /əʊ/ to [u] among the basilectal speakers registers 85 % as opposed to 35 % among the acrolectal. So, the features that do not move towards CamE and RP later in education are effectively stable ethnolectal features. As speakers advance in education so too do they lose some of the ethnolectal features they used in primary school located in the village.

In the case of WimE , the diphthong tested was /eɪ/, and as illustrated in Table 6.5, it was not only reduced or simplified to a monophthong but was also moved one level lower than its individual components, i.e. to /ɛ/. Just as with NsoE above, the basilectal speakers produce more ethnolectal patterns than the acrolectal: 90 % as opposed to 65 %. Again, educational advancement and exposure could be accountable for this.

Table 6.5 Wimbum English: able→ [ε]ble (n = 40)

The results in Table 6.5 show an average of 77.5 % of the 40 WimE respondents’ realisation of the characteristic WimE [ɛ] in [ɛ]ble for RP /eɪ/ble and CamE /e/ble. In WimE, the RP diphthong /eɪ/ has not only been reduced to a monophthong as in CamE, it has also dropped one step lower than in CamE . The diphthong /eɪ/ is not attested in Limbum, the L1; so, this is also potentially a case of substrate influence where speakers substitute an unfamiliar sound by a familiar one, in this case /ɛ/. In Limbum, /ɛ/ is a recurrent vowel, which also observes length as in the words lɛɛ (bat), tɛɛ (stand), sɛɛ (slaughter), and wɛɛ (hunting) (see Tamfu 1999, p. 19) .

Curiously, WimE does not replace the RP diphthong /eɪ/ by the CamE /e/ which is also a Limbum phoneme seen in words like bep (bad), ye (eat), wep (bitter leaves), and be(invite) (see Tamfu 1999, p. 17) . It rather prefers the phoneme /ɛ/, perhaps because of its position at word beginning, since /ɛ/ is more open than /e/. Additionally, because /ɛ/ is more spread, it requires less articulation effort than /e/, and is, therefore, easier to realise especially in a second language. Here, we can talk of the principle of least effort (Wells 1982) , which, as attested in the literature, is a common feature in the speech of second or foreign language speakers or learners.

6.2 Vowel Shortening Plus Lowering

The process of vowel shortening is also common in indigenised varieties of English. It generally occurs in closed syllables, although some varieties also realise it in open syllables. Sometimes, shortening is followed by another process, e.g. vowel lowering , in which the reduced vowel is actually realised as another vowel, lower in position of articulation in the mouth (Table 6.6).

Table 6.6 Nso English: meal → m[ε]l (n = 39)

The vowel targeted for this process in NsoE was the long, high front vowel /i:/ in the word ‘meal’. The basilectal respondents realised it in two ways: reducing it to a lower vowel [ɛ] (65 %) or diphthongising it into [ie] (5 %).⁶ The tertiary speakers instead preferred only [ɛ] (55 %). It could be assumed that the [ie] realisation disappears as the speakers advance in education and are exposed to other varieties of the language.

The results in Table 6.6 show that as high as 60 % of the NsoE informants shorten the RP long /i:/ but realise it a level lower as [ɛ] yielding m[ɛ]l. CamE only shortens it to the short counterpart /i/ but does not lower it. From the example word tested, i.e. ‘meal’, it can be said that vowel lowering in NsoE occurs mostly after nasals.⁷

Again, education seems to be a deciding factor here: the 20 % realisation of RP /i:/ is attested exclusively among the acrolectal speakers . CamE realisation increased by 5 % between the basilectal (20 %) and the acrolectal (25 %).

6.3 Vowel Lowering

Reference has been made above to vowel lowering as a type of variation in ethnolectal accents in Cameroon. Lowering involves the replacement of a sound with another sound that is produced at a lower point in the mouth. The example chosen to illustrate this is from WimE . The RP short front vowel /ɪ/ as in the word ‘promise’ was tested in the interviews. Gathering from Table 6.7, 55 % of the WimE speakers tested lowered /ɪ/ to /ɛ/.

Table 6.7 Wimbum English: promise → prom[ε]se (n = 40)

Interestingly, the realisation of the WimE accent feature [ɛ] here is not in a diphthongal space but in a monophthongal space. RP /ɪ/ is absent from both CamE and Limbum phonetic inventories. This is, apparently, one of the reasons why RP /ɪ/ is raised to [i] in CamE, while in WimE, CamE /i/ is moved two steps lower to [ɛ]. Unlike in Lamnso’, in which the structure <mi> is rare, <mi> is permissible in Limbum as in mì(finish) and tàami (go across). In spite of this <mi> permissibility, the WimE speakers still lowered /i/ to [ɛ] realising RP prom[ɪ]se and CamE prom[i]se as prom[ɛ]se. Again, as in most of the examples above, the place of the substrate seems pervasive. It is natural to realise /ε/ in this position because of its dominance in Limbum.

In summary, Table 6.8 captures the major ethnolinguistic variation in CamE pronunciation. It is not based on the discussion above alone but also on other features identified by researchers, some of them presented in Sect. 6.3 above. These features, as the last two rows show, include consonants as well.

Table 6.8 Phonological features highlighting heterogeneity in CamE

From Table 6.8 it can be deduced that there is significant ethnolinguistic dissimilarity as much as there is similarity in these ethnolects . The shaded portions show exactly where the variation is.

One feature that is shared by a number of ethnolects is the [u] for RP /əʊ/ and CamE /o/ phenomenon. It is shared by Nso’, Kom, and Wimbum ethnic accents, although it is most recurrent in NsoE . Generally, Bafut speakers of English realise the CamE /i/ and RP /ɪ/ as [ɨ], suggesting that the Bafut language has a clipped vowel system which is transferred unto English. The NsoE and WimE speakers surveyed generally have [ɛ] in this position, indicating that they realise the CamE front vowel /i/ two steps lower when it occurs after a nasal consonant, but the recurrence of /i/ and /ε/ after nasal consonants alternates or freely varies more in NsoE than in WimE. In Lamnso’ /i/ and /ε/ alternate between nasals and after nasals. When Lamnso’ speakers transfer this alternation process to English, there is the tendency to lower RP /eɪ/ and CamE /e/ to [ɛ] after nasals, or to raise them to [i] between nasals in NsoE.

7 Conclusion

This chapter set out to identify and illustrate salient vocalic features of two ethnolects in CamE : NsoE and WimE . It focused on three main processes, namely, diphthong reduction or simplification, vowel shortening , and vowel lowering . Using data collected through interviews, the chapter has shown that the major source of the vocalic variations identified is the substrate influence from indigenous Cameroonian languages acquired as L1 by the respondents. We could also add the principle of least effort which facilitates the choice of certain sounds in given phonetic environments. However, this choice is not always the same across ethnolects. While some features may be similar across ethnolects, or with mainstream CamE, others tend to be either specific to particular ethnolects or are realised in specific phonetic environments.

As far as the impact of education and exposure to other varieties are concerned, it has been shown that the higher respondents go in education the lower the amount of ethnolectal pronunciation features they use. The results in Tables 6.4, 6.5, 6.6, and 6.7 suggest clearly that the recurrence of ethnic features reduces significantly between the basilectal and acrolectal speakers . Conversely, the recurrence of CamE and RP realisations rather increase from the basilectal to the acrolectal. In some cases, the basilectal speakers do not even realise RP sounds, but these are realised by the acrolectal, e.g. m[i:]l → m[ε]l (meal) in Table 6.6 among the NsoE speakers where the RP variant m/i:/l is used by 20 % of the 39 respondents, all of them acrolectal speakers. This notwithstanding, education and exposure have not prevented ethnolectal features from surfacing in the speech of acrolectal speakers , sometimes reaching as high as 77.5 % (Table 6.5) or 60 % (Tables 6.4 and 6.6). What this implies is that, the features that do not disappear can effectively be treated as authentic markers of these ethnolects .

We could also factor in the sociolinguistic setting and how it possibly facilitates the retention of ethnolectal features. Given that the Nso’ and Mbum areas make sustained use of English besides their L1s and other languages, and that the teaching of English in schools orients speakers towards a target norm, substrate features will continue to thrive in the ethnolects spoken there. Gut’s (2007) norm-orientation hypothesis could be used to account for this. This means, therefore, that it is the predominant ethnic languages used as L1, whose features determine the way in which these multilingual speakers articulate English. Most of the results reported on above seem to testify to this.

The chapter is only the tip of the iceberg for a country that has more than 250 indigenous languages . Further research on both segmental and supra-segmental features, vocalic and consonantal features, and from synchronic and diachronic perspectives will certainly throw more light on ethnolects and their relationship to mainstream (national) varieties of English in both non-native and native communities, especially given the current spread of the language.