Pronunciation model: East African English

View the key for East African English here.

There is little clarity in the distinction between North, Central, North-East, Southern, and East African regions. However, following Schmied (2008) and in-keeping with Oxford Dictionaries’ definition of East Africa, the focus here is on the Eastern nations of Kenya, Uganda, and Tanzania, which share a common linguistic background. However, there is widespread variation and the patterns below are tendencies rather than rules. There are also important differences between national and subnational standards, particularly in Kenya and Uganda where subnational variation is prominent; few of the subnational standards are actively followed in a systematic way, others being widely tolerated and only a few rejected as ‘substandard’.

Useful summaries of the history of East African English (EAfE) are given by Schmied (1991, 2008, 2015) and interested readers are directed to his work, reflecting on the impacts of colonial administration, Christian missionaries, and railway construction. The colonial language policy was quite complex, an essentially trilingual setup valuing local languages (for local communication and basic education), the African lingua franca (usually Kiswahili, for cross-culture interactions), and English (for administrative, legal, and more advanced educational purposes). The forms of English spoken in modern East Africa are also complex to describe, heavily influenced by a speaker’s formal education and their occupation (impacting the necessity they have for speaking English). Some speakers will be native English speakers, though fewer than in Southern or West Africa, while most will speak it as a second language.

Sociolinguistically, the East African nations differ as to the status of Kiswahili and the changes in the official status of English since independence in the 1960s. It is also difficult to establish knowledge and use of English across the region in the absence of reliable data. Attitudinally, it is only in pronunciation that there is widespread acceptance of deviation from British English, with acceptance of Standard English syntax and grammar spoken with African accents. Yet African forms differ from the British ‘standard’ for a variety of reasons, including interference from the local native language, general language learning strategies, and written English exposure (since the written form has authority over the spoken form, leading to articulations such as word-initial /h/ in heir). Pronunciations in EAfE include many ‘nonstandard’ features, the main ones being listed below. The OED EAfE model is shaped by the pronunciations of our consultant speakers as well as key literature.

EAfE is discussed as claiming only five distinct vowel qualities in contrast to seven-vowel West African varieties. The table below is an adaption of Schmied’s tables from his 2008 and 2015 descriptions, with a key systemic alteration.

East African vowels

KeywordSymbolKeywordSymbolKeywordSymbolKeywordSymbol
KITiFLEECEiNEARiaHAPPYi
DRESSeGOOSEuSQUAREeaLETTERa
TRAPaPALMaCUREuaRABBITi
BATHaSTARTaFACEeADDEDi, e
LOToNURSEa, ePRIDEaiBEAUTIFULu
CLOTHoNORTHoVOICEoiPIANOi
STRUTaFORCEoMOUTHauAGOa
FOOTuTHOUGHToGOAToBECAUSEi

EAfE does not have a consistent long-short vowel distinction. The equivalents of the British English ‘short’ vowels are also closer to the edges of the vowel space, and for this reason kit seems better reflected as /i/ rather than /ɪ/, although Schmied uses the latter in his 2015 discussion. Similarly, although Schmied transcribes the pride vowel as /aɪ/, the second element does not seem to be significantly different from the kit vowel and Bobda (2001) transcribes tribal as /traibo/ or /traibol/. Indeed, some traditionally diphthong vowels are either pronounced with a barely-perceptible second element (as in face) or tend towards being produced as two consecutive monophthong vowels (as in voice). For simplicity this latter phenomenon is only reflected in the audio in OED and for stress marking purposes usually treated as a single syllable. Many British vowel distinctions are not held in EAfE, for example, strut-palm-nurse (though nurse is particularly variable according to spelling and/or region). Schmied notes that a consequence of this is that vowels in full syllables are therefore underdifferentiated. In contrast, unstressed syllables may be overdifferentiated (policeman and policemen not being homophonous).

The EAfE phonemic vowel spaces are perceptually larger than for many other varieties of English. In the OED audio recordings, /e/ is sometimes a crisp close [e], elsewhere a more open [ɛ]. /i/ similarly spans qualities from [i] to [ɪ], /a/ can sound more open-central than open-front, and /o/ spans a range of rounded back vowel heights (though maintaining distinction with /u/).

/r/ before vowels may be trilled or flapped. EAfE as a rule is non-rhotic, but evidence from OED’s speakers suggests that post-vocalic /r/ may be pronounced if the root language would potentially indicate it, as in Turkana.

There several consonant conflations observed within different groups across the region but few consistently. For example, the stigmatized /l/-/r/ merger is widespread, but may resolve differently depending on the ethnic group of the speaker. For this reason the merger is not reflected here. British English /θ/ and /ð/ are sometimes EAfE /t/ and /d/ depending on ethnic group and education, although this is stigmatized and other realizations are found. For many speakers, there is a form of ‘partially voiced’ /ð/ instead of distinct /θ/ and /ð/. For consistency, OED gives only the /ð/ EAfE pronunciations, though the audio may reflect more variable voicing levels. The influence of African languages may result in inserted or hypercorrection-deleted nasal consonants before stop consonants, but these are evaluated case-by-case.

The presence of syllabic consonants is not discussed by much EAfE literature, yet they are a salient feature of the OED EAfE audio and pronunciation system. Syllabic /m/, /n/, and /ŋ/ are observed, and potentially /l/ or even /r/, though distributions may differ from those of other English varieties. Syllabic nasals are regularly found in borrowed but frequent words where a nasal is immediately followed by an obstruent (plosive, fricative, or affricate). In some cases where a syllabic consonant could be found in British English (e.g. Bible), there is a clear vowel in EAfE pronunciations, most commonly of [o] or [ɔ] quality and reflected in OED as /o/. Some instances in EAfE will be more /u/ as in the OED West African English model and there is not a clear distinction between the two varieties in this regard, but for the purpose of offering consistent and practical pronunciation models, OED retains the vowel qualities most commonly used in the literature on each (see also Wolf 2010). Also note that some word-final /l/ qualities may be entirely vocalized, transcribed in OED as e.g. tribal /ˈtraibo(l)/.

Root-language voiced palatal plosives and nasals may be borrowed into a code-mixed realisation by many speakers, but are reflected in EAfE as /dʒ/ and /nj/ respectively. The voiceless velar fricative /x/ may be found in some words borrowed from Swahili transliterated with kh.

Given the preference of many African languages towards a consonant-vowel syllable structure, consonant clusters are sometimes dispreferred, resulting in epenthesis (adding a vowel to make another syllable) or reduction (losing one of the sounds). Although highly variable between speakers, final consonant clusters may have the final consonant dropped (/neks/ for next, /han/ for hand). Some word-final vowel epenthesis is also noted by Bobda, and this is assessed on a case-by-case basis.The general preference for epenthesis over elision extends to phrase-level, e.g. ‘and go’ more likely to have epenthetic schwa between /d/ and /ɡ/ than the West African English tendency towards elision of the /d/ (Wolf 2010). OED’s editors consider a more technical description of common patterns based on Schmied (2008) as to how entries are most likely to be realized, though transcriptions should be simply regarded as reflecting one possible realization rather than necessarily the most frequent.

Following existing OED pronunciation models, stress placement within each entry reflects how it would be perceived by knowledgeable British and American speakers, rather than necessarily the instinctive self-perceived placement of EAfE speakers. This approach therefore does not always agree with e.g. documented Kiswahili prosodic patterns or syllable divisions, but enables OED to highlight the core differences and similarities as to how different varieties of English compare at a perceptual level amongst English speakers (though the way ‘pitch’ is used in different varieties of English is the primary complicating factor).

Schmied describes the EAfE tendency towards syllable-timed rhythm. In some of the OED models (e.g. Singapore & Malaysian English), this is shown in the OED by a ‘flattened’ stress structure, with nearly all syllables transcribed as having at least secondary stress marks. However, the OED EAfE speakers maintain a clear distinction in each word between the primary-stressed syllable and other syllables, and unstressed syllables are readily perceived as such even though the vowel qualities are not significantly reduced. EAfE is therefore treated more similarly to West African English in this regard.

Stress patterns themselves tend to be more regularized than in British or American Englishes, though there are occasions where multiple stress patterns are found. Although highly variable, there is a tendency towards putting final word stress on suffixes such as –ize and -ate except for where the frequency and familiarity are such that the British pattern is more commonly adopted. The use of stress to make noun-verb distinctions is shown, although listeners may frequently encounter speakers who do not maintain this.

Key sources for East African English