South African English phonology

• /iː/ is a long close front monophthong [iː], either close to cardinal [iː] or slightly mid-centralised. It does not have a tendency to diphthongise, which distinguishes SAE from Australian and New Zealand English.[4]
• The HAPPY vowel is phonemically equal to /iː/ in General and Broad and to /ɪ/ in Cultivated. Thus, happy and immediately are phonemically /ˈhæpiː/ and /əˈmiːdiːətliː/ in General and Broad and /ˈhæpɪ, ɪˈmiːdɪətlɪ/ in Cultivated. The pronunciation with FLEECE is sometimes considered to bare secondary stress, so that happy can be analysed as /ˈhæpˌiː/.[5]
• /ɵ/ is typically a weakly rounded retracted central vowel [ɵ̠], somewhat more central than the traditional RP value. Younger speakers of the General variety (especially females) often use a fully central [ɵ]. Backer and sometimes more rounded variants ([ʊ ~ ʊ̹]) occur before [ɫ]. Broad SAE can feature a more rounded vowel, but that is more common in Afrikaans English.[4][6][7]
• /ʉː/ is usually central [ʉː] or somewhat fronter in White varieties, though in the Cultivated variety, it is closer to [uː] (typically not fully back, thus [u̟ː]), which is also the normal realisation before [ɫ] in other varieties. Younger (particularly female) speakers of the General variety use an even more front vowel [yː], so that food [fyːd] may be distinguished from feed [fiːd] only by rounding. The vowel is often a monophthong, but there is some tendency to diphthongise it before sonorants (as in wounded [ˈwʉundəd] and school [skʉuɫ]).[8][9]
• /e/ is close-mid [e] or higher [e̝] in General, often with centralisation [ë ~ ë̝]. Closer variants are typical of female speech. General /e/ is similar enough to /ɪ/ in RP and similar accents as to cause perceptual problems for outsiders. Broad variants are very similar to the General ones, but in Cultivated the vowel can be as open as [e̞] (within the RP norm). In General and Broad, the vowel can be lowered to [ɛ] or even [æ] when it occurs before [ɫ].[10][11][12]
• /eː/ is realised as a centring diphthong [ɛə] in the Cultivated variety. The broad realisation is a monophthong, close-mid [eː] or higher [e̝ː] (but not as close as to merge with /iː/). In General the vowel is often a close-mid monophthong (especially for younger speakers), but it can also be open-mid [ɛː] or the Cultivated variant [ɛə]. Monophthongal realisations are often stigmatised by General speakers, even those who diphthongise. When realised as [eː], /eː/ can be considered to be the long counterpart of /e/ as there is almost no difference in quality between the two.[10][13][14]
• In General and Broad, /øː/ is a rounded front vowel that varies between close-mid [øː] and open-mid [œː], with the former being more usual.[15] In Cultivated an unrounded, RP-like mid central [ɜː] is used, with a more open [ɐː] being a possible alternative, though the latter could also be thought of as a hypercorrection.[16]
• In Cultivated, final /ə/ is sometimes more open than in other varieties ([ɐ] or even [ä]), so that comma can be pronounced [ˈkɒmɐ] or [ˈkɒmä] instead of the more usual [ˈkɒmə].[5]
• /oː/ is close-mid [oː], as in Australian and New Zealand English. A somewhat more open realisation [o̞ː] is possible in the General variety and usual in Cultivated SAE.[8][9]
• In Cultivated and General, /æ/ is realised as [æ] or somewhat closer ([æ̝]). However, the new prestige value in younger Johannesburg speakers of the General variety (particularly those who live in the wealthy northern suburbs) seems to be open front [a], the same as in Modern RP. Before [ɫ], the fully open [a] is the norm in the General variety, whereas before voiced stops and nasals (except the velar /ŋ/) the vowel tends to be centralised and lengthened to [æ̈ː], often with slight diphthongisation ([æ̈ːə]). Broad /æ/ can be as close as mid [ɛ̝], encroaching on the Cultivated realisation of /e/.[10][11][12][17]
• /ɐ/ is normally an open vowel, either central [ɐ̞] or retracted front [a̠], although younger, especially female speakers of the General variety may realise this vowel as close and front as open-mid centralised front [ɛ̈].[12]
• /ɒ/ is realised as a mid-centralised cardinal [ɒ] with very weak rounding ([ɒ̜̽]), or sometimes unrounded [ɑ̽]. The unrounded variant can be raised to open-mid [ʌ̈], especially when it occurs in non-prominent syllables. The unrounded open-mid variant is gaining ground among General speakers in Cape Town and Natal, who generalise it to other positions (so that lot becomes [lʌt] instead of the traditional [lɒt ~ lɑt]). There have been no reports of LOT merging with STRUT.[4][12]
• In General, /ɑː/ is an open back unrounded vowel [ɑː], either fully back (as in Broad) or slightly more front. The broad variants can be rounded [ɒː], sometimes with raising to [ɔː] and/or shortening to [ɒ] or [ɔ], so that Broad dark [dɒk] can encroach on General and Cultivated dock [dɒk], or at least be differentiated from the latter by length. Weak rounding of this vowel is now somewhat less stigmatised than it was in the 20th century. In Cultivated, a categorically unrounded advanced back [ɑ̟ː] or (more rarely) central [äː] is used.[7][18]

In General and Broad, this vowel is split between a more front realisation ([ɪ] in General, [i] in Broad) and a more central realisation ([ɪ̈] in General, [ɪ̈] or [ə] in Broad). More front variants are used in contact with velar and palatal consonants, whereas more central realisations are used in other environments (and therefore are more common), with an even more retracted [ɯ̽] being possible before the velarised allophone of /l/, so that pill (phonemically /pəl/) can be realised as [pɯ̽ɫ], close to pull [pʊɫ]. The Cultivated variety lacks this split, and uses a lax front [ɪ] in every position.[19]

John Wells analyses the front variants as /ɪ/ and the central variants as /ə/, which makes /ə/ one of the stressable vowels in South African English. Some scholars prefer to analyse all of the variants as allophones of /ɪ/.[20] This article adopts the former approach, even though the split does not create any minimal pairs.

• /eɪ/ is realised as [eɪ] in General and Cultivated. In General, more open onsets [ɛɪ ~ æɪ] are possible in accents closer to Broad SAE. However, in Broad SAE, the first element is actually back [ʌɪ].[21]
• /aɪ/ is a diphthong [aɪ] in Cultivated. In General and Broad, it is more often a monophthong [aː], with a diphthong with a retracted onset [ɑ̽ɪ] also possible in Broad.[21]
• /ɔɪ/ is realised as [ɔɪ] in all White varieties. In the Cultivated variety, the first element may be lowered to [ɒ].[21]
• /œɨ/ regularly has a rather front onset in Cultivated and General varieties. In Cultivated, it is realised as [ɛʊ ~ œʊ], with the former being more common. In General, the onset is always rounded, and the offset is central [œʉ ~ œɨ̞], with a tendency to monophthongise it to [œː]. The Broad realisation is back [ʌʊ].[21]
• /aɤ/ is realised as [ɑ̈ʊ] in Cultivated, [ɑː] in General and [æʊ] in Broad.[21]
• /ɪə/ is realised as [ɪə] in White varieties, but Broad speakers tend to monophthongise it to [ɪː], especially after /j/.[21]
• /ʉə/ is a glide from [ɵ̠] to [ə] in the Cultivated variety, whereas in Broad it is monophthongised to [oː]. The General variety varies between these two, with a growing tendency to use the monophthong [oː] (the same as the THOUGHT vowel or somewhat lower), especially when /j/ doesn't precede.[22]

• In Broad White South African English, voiceless plosives tend to be unaspirated in all positions, which serves as a marker of this subvariety. This is usually thought to be an Afrikaans influence.[23][24]
• General and Cultivated varieties aspirate /p, t, k/ before a stressed syllable, unless they are followed by an /s/ within the same syllable.[23][24]
  • Speakers of the General variety can strongly affricate the syllable-final /t/ to [ts], so that wanting /ˈwɒntɪŋ/ can be pronounced [ˈwɒntsɪŋ].[25]
• /t, d/ are normally alveolar. In the Broad variety, they tend to be dental [t̪, d̪]. This pronunciation also occurs in older speakers of the Jewish subvariety of General SAE.[23][24]

• /x/ occurs only in words borrowed from Afrikaans and Khoisan, such as gogga /ˈxoxa/ 'insect'. Many speakers realise /x/ as uvular [χ], a sound which is more common in Afrikaans.[23]
• /θ/ may be realised as [f] in Broad varieties (see Th-fronting), but it is more accurate to say that it is a feature of Afrikaans English. This is especially common word-finally (as in myth [mɪ̈f]).[23][24]
• In Indian variety, the labiodental fricatives /f, v/ are realised without audible friction, i.e. as approximants [ʋ̥, ʋ].[26]
• In General and Cultivated varieties, intervocalic /h/ may be voiced, so that ahead can be pronounced [əˈɦed].[27]
• There is not a full agreement about the voicing of /h/ in Broad varieties:
  • Lass (2002) states that:
    • Voiced [ɦ] is the normal realisation of /h/ in Broad varieties.[27]
    • It is often deleted, e.g. in word-initial stressed syllables (as in house), but at least as often, it is pronounced even if it seems deleted. The vowel that follows the [ɦ] allophone in the word-initial syllable often carries a low or low rising tone, which, in rapid speech, can be the only trace of the deleted /h/. That creates potentially minimal tonal pairs like oh (neutral [ʌʊ˧] or high falling [ʌʊ˦˥˩], phonemically /œɨ/) vs. hoe (low [ʌʊ˨] or low rising [ʌʊ˩˨], phonemically /hœɨ/).[27]
  • Bowerman (2004) states that in Broad varieties close to Afrikaans English, /h/ is voiced [ɦ] before a stressed vowel.[23]

• General and Broad varieties have a wine–whine merger. However, some speakers of Cultivated SAE (particularly the elderly) still distinguish /hw/ from /w/, so that which /hwɪtʃ/ is not homophonous with witch /wɪtʃ/.[28][29]
• /l/ has two allophones:
  • Clear (neutral or somewhat palatalised) [l] in syllable-initial and intervocalic positions (as in look [lɵk] and polar [ˈpœɨlə]).[28][29]
    • In Cultivated variety, clear [l] is often also used word-finally when another word begins with a vowel (as in call up [koːl ɐp], which in General and Broad is pronounced [koːɫ ɐp]).[28][29]
  • Velarised [lˠ] (or uvularised [lʶ]) in pre-consonantal and word-final positions.[28][29]
    • One source states that the dark /l/ has a "hollow pharyngealised" quality [lˤ],[25] rather than velarised or uvularised.
• In the Broad variety, the sequences /ən/ and /əl/ tend not to form syllabic [n̩] and [l̩], so that button /ˈbɐtən/ and middle /ˈmədəl/ are phonetically [ˈbɐtən] and [ˈmɪ̈dɯl] (compare General [ˈbɐtn̩] and [ˈmɪ̈dl̩]). John Wells analyses the broad pronunciation of these words as having a secondarily stressed schwa in the last syllable: /ˈbɐtˌən/, /ˈmədˌəl/.[10]
• In Cultivated and General varieties, /r/ is an approximant, usually postalveolar or (less commonly) retroflex. In emphatic speech, Cultivated speakers may realise /r/ as a (often long) trill [r]. Older speakers of the Cultivated variety may realise intervocalic /r/ as a tap [ɾ] (as in very [ˈveɾɪ]), a feature which is becoming increasingly rare.[28][30]
• Broad SAE realises /r/ as a tap [ɾ], sometimes even as a trill [r] - a pronunciation which is at times stigmatised as a marker of this variety. The trill [r] is more commonly considered a feature of the second language Afrikaans English variety.[28][29]
• Another possible realisation of /r/ is uvular trill [ʀ], which has been reported to occur in the Cape Flats dialect.[31]
• South African English is non-rhotic, except for some Broad varieties spoken in the Cape Province (typically in -er suffixes, as in writer [ˈraɪtɚ]). It appears that postvocalic /r/ is entering the speech of younger people under the influence of American English.[28][29]
• Linking /r/ (as in for a while /foːr ə waɪl/) is used only by some speakers.[28]
• There is not a full agreement about intrusive /r/ (as in law and order) in South African English:
  • Lass (2002) states that it is rare, and some speakers with linking /r/ never use the intrusive /r/.[27]
  • Bowerman (2004) states that it is absent from this variety.[28]
• In contexts where many British and Australian accents use the intrusive /r/, speakers of South African English who do not use the intrusive /r/ create an intervocalic hiatus. In these varieties, phrases such as law and order can be subject to the following processes:[28]
  • Vowel deletion: [ˈloːn ˈoːdə];[28]
  • Adding a semivowel corresponding to the preceding vowel: [ˈloːwən ˈoːdə];[28]
  • Inserting a glottal stop: [ˈloːʔən ˈoːdə]. This is typical of Broad varieties.[28]
• Before a high front vowel, /j/ is foritified to [ɣ] in Broad and some of the General varieties, so that yeast can be pronounced [ɣiːst].[28]