Pronunciation model: Singapore and Malaysian English

View the key for Singapore and Malaysian English here.

In many of the varieties now included in the OED, the inclusion/exclusion of particular features is subject to ongoing review, but the entire concept of the variety described here as Singapore & Malaysian English (‘SME’) will itself require review over the coming years. Baskaran (2008) cites a description by Tongue in 1974, who acknowledged that his own description of the English of ‘Singapore and Malaysia’ would not be appropriate within a century given the increasing diversity. Baskaran notes the political disconnection of Singapore and Malaysia since 1965 and resultant sociolinguistic variation, in addition to the different language policies of the countries, and highlights the linguistic substrate differences such that the Chinese grouping is in the majority in Singapore but a minority in Malaysia. It is only within the past decade, however, that Malaysian English has gained traction as a variety of study and the immense variation reported thus far (at different levels of proficiency and in different geographical locations) means that the creation of a truly encompassing yet distinct ‘Malaysian English’ model would not be feasible (Low & Tan 2016). The SME model proposed, therefore, is based predominantly on a Colloquial Singapore English (‘CSE’) model described by Wee (2008), influenced by Wells (1982) and Low (2012) and, in the context of part of Low & Tan’s (2016) discussion, focusing on those features most consistent with mesolectal Malaysian English [MalE] similar to Baskaran’s (2008) preliminary analysis and that of Hashim & Tan (2012).  A number of additional sources have been consulted on specific aspects of SME pronunciation. First, however, it is valuable to consider the linguistic situations within Singapore and Malaysia.

Singapore English is inevitably influenced by the other three official languages: Malay (the National language, primarily ceremonial or military), Mandarin Chinese, and Tamil, with 74% of the population Chinese, 13% Malay and 9% Indian (2010 census). Wee (2008) explains how English is intended by the government to be a neutral language, but there is a difference and tension between the standard variety and its colloquial counterpart (commonly also known as Singlish). The latter is viewed by the state as potentially undermining the development of proficiency in the standard (with economic implications). Singapore English can be viewed as a lectal continuum or diglossia, yet neither approach appears satisfactory. While it is potentially communicative choice to use the standard over colloquial form, rather than a matter of proficiency (the diglossic approach), elements of the varieties cross over rather than remaining entirely distinct (see Wee 2008). In any case, ‘[m]ost of the phonetic characteristics of Singaporean English are obviously to be attributed to interference from the Chinese phonological common core (inc. limitations on the range of possible syllable-final consonants)’ (Wells 1982, pp.645-646).

In contrast, around 67% of the Malaysian population is indigenous Malay (Austronesian speakers) and Austroasiatic counterparts (aboriginal tribes), with around 25% from (predominantly Mandarin) Chinese settlers, and around 7% Indian (2010 census).  The official national language for education is Bahasa Malaysia, coexisting with English as official languages until 1967, English being a ‘strong second language’ ever since (Baskaran 2008). Generally the status of English warrants its use in meetings, conferences and liaison with international audiences, as well as the second language in all schools and some university courses. Baskaran (2008) discusses a lectal system to MalE, the acrolect tending towards a British standard, a basilectal patois form, with the mesolect between them being the informal style used among Malaysians.


Several vowel features of SME are particularly salient, features described for both CSE and MalE. Firstly, there is a difference between the vowel in CURE on the one hand and the diphthong appearing in poor, tour and sure on the other. Secondly, the vowel in AGO is simply indicative of there being unreduced vowel in pretonic position (that is, before the primary stressed syllable; this pattern is found elsewhere, but not as consistently). A stressed /ə/ is possible in NURSE, while the STRUT vowel is further back (but Low’s experimental CSE research suggests it may be further forward and Hashim & Tan suggest the MalE start/strut/palm conflation may be nearer /a/). Perhaps most significantly, there is conflation of FLEECE and KIT, and GOOSE and FOOT (as in HKE), but also DRESS and TRAP, and PALM and STRUT (only the FLEECE-KIT and DRESS-TRAP conflations were identified in Low’s acoustic research, but the OED model is more concerned with perceptual contrasts). Various sources use different vowels for DRESS and TRAP in Singapore English (supporting Low’s findings that speakers distinguish between the two in citation form but this is lost when recorded in conversation), but to adopt these symbols would lose the widely accepted conflation of these vowels.

Further, the British English long vowels are short (no length marking), hypothesized for MalE to be under the influence of Bahasa Malaysia which also lacks long vowels, and several vowels which are diphthongs in other varieties are monophthongs here (e.g. SQUARE, GOAT). Where /i/ precedes another vowel in the same syllable, it is rendered as /j/ to avoid the impression of two syllables (e.g. teow as /tjau/ rather than /tiau/), even if the usual gliding is minimal. An additional small point, but particularly noted in MalE, is that /i/ is not diphthongized to NEAR before /r/ in words such as serious.

SME as transcribed here is non-rhotic and without /t/-flapping (though there is evidence for the Americanization of Asian English phonologies, including Singapore English; see accompanying chapters to Low & Tan 2016).

The influence of neighbouring languages has resulted in SME having a more syllable-timed rhythm than many other varieties, with each syllable having a similar duration. In addition, loudness is the main differentiating factor for stress, not pitch, with intonation within a word usually level except for a few particles. This can be heard in the spoken pronunciations of each entry. Levels of stress are therefore perceived as less distinct than in British or American Englishes. This is shown in the OED by a ‘flattened’ stress structure, with nearly all syllables transcribed as having at least secondary stress marks and with double-primaries permitted on multi-word entries. Placement may also differ from British/American but is evaluated on a case-by-case basis. The HKE final-element stressing of compounds is also a feature here, as is the MalE lack of noun-verb stress contrast. The strong-weak distinction of British English (as heard in words such as the, a, of, and in connected speech) is not common in SME. Syllabic consonants are not represented (always with a vowel preceding the consonant instead, as in Wee 2008), possibly influenced by the CVC syllable structure of Bahasa Malaysia, and as in HKE, liaison (the ‘linking’ or ‘intrusive’ /r/ as in British near it /nɪər ɪt/) is not commonly found outside of particular acrolectal MalE speech so is also not a feature of the SME model.

Dental sounds /θ/ and /ð/ become stopped to /t/ and /d/ respectively, except for final /ð/ becoming /t/ (this is the approach taken here, despite some disagreement as to the precise articulation). Word-final /b/, /d/, /ɡ/, /z/, /ʒ/ and /dʒ/ are represented as their voiceless equivalents. Word-final stops are deleted if they are the second element of a final CC consonant cluster, even if they are formed by the addition of an inflection such as -ed (e.g. link /liŋ/, backed /bɛk/). They are retained by some speakers, however, if the preceding consonant is /l/ (so help and helped are both /hɛlp/) and this is the approach adopted here. There is a two-consonant limit in syllable coda clusters medially and finally – there seems to be slightly less of a restriction in CSE than in MalE but the greater restriction is followed in the OED. Word-final stops usually have no audible release and are often glottalized, but this is only reflected in the spoken pronunciations (not the transcriptions).

Several features were considered for inclusion but set aside at the present time. CSE’s tendency to make word-final /θ/ (and by extension /ð/) into /f/ (as in filth, breathe) is not notably documented in MalE. The CSE metathesis (switching) of /sp/ as /ps/ is also not currently noted in MalE so does not form part of the model. Similarly, Wells (for Singapore English) and Baskaran (for MalE) both discuss cluster reductions involving /l/, but the mechanics are too variable to present a coherent and consistent approach across the varieties. Wells also notes that those with the greatest influence from their first language may also have a lack of /s/-/ʃ/ distinction, or not distinguish /r/-/l/ (if from the Chinese population), /f/-/p/ (if Malay), or /v/-/w/ (if Indian). Short vowel lengthening in MalE is also not represented here, nor is word-final stop glottalization, nor particular voicing features only found in an unknown proportion of MalE speakers.