The Sounds of Mauritian Creole: Phonology and Pronunciation

Mauritian Creole, or Kreol Morisien, has a smaller sound system than French, and the differences are systematic rather than random. This page sets out the phoneme inventory, the processes that shaped it, and how it maps onto the standard 2011 orthography.

Vowels

Five oral vowels and three nasal.

Nasal: /ẽ/, /õ/, /ã/.

Two things to note straight away.

The mid vowels have open and close allophones, conditioned by syllable type. /e/ is realised as [ɛ] in closed syllables and [e] in open ones, and /o/ behaves the same way, as [ɔ] and [o]. So [e] and [ɛ] are not separate phonemes here, which is one of the clearest reductions relative to French.

The nasal vowels are written with close symbols but pronounced mid-low. /ẽ/ and /õ/ are "always given mid-low pronunciations", [ɛ̃] and sometimes [æ̃] for the first, [ɔ̃] for the second. The phonological labels and the phonetic values come apart, which matters if you are comparing transcriptions across sources.

The emerging sixth vowel

Schwa is becoming established, and it is the one clearly register-sensitive vowel in the language. It is "not heard in basilectal speech but is becoming frequent in other varieties, particularly in words which traditionally had /i/ deriving from schwa in their initial syllable". The standard example is the word for water: basilectal dilo, with a schwa in other varieties.

The 2011 orthography provides no grapheme for schwa, so this variation is invisible in writing.

What happened to the French vowels

French has at least seven peripheral vowels; Mauritian Creole has five, with no mid-high and mid-low contrast. The French front rounded vowels were unrounded, /y/ going to /i/ and /ø/ and /œ/ to /e/. None of the immigrant languages of Mauritius had front rounded vowels either, which is part of why this change is often discussed in substrate terms, though that attribution is a position rather than a consensus.

Denasalisation, the process that sets the language apart

This is the headline finding and it is worth stating plainly, because it is frequently reported backwards. "Denasalization of historically nasalized vowels distinguishes Mauritian Creole from other French creoles." The language did not add nasality under African influence. It lost nasality that French had.

It happens in two environments.

Word-finally, where a nasal vowel was followed by a voiced plosive, the vowel denasalises and a nasal consonant appears:

• ansam, "together", from ensemble
• kokom, "cucumber", from concombre
• vann and vande, "sell", from vendre

Word-finally after /m/:

• lame, "hand", from la main
• pima, "chilli", from piment

The doubled n of the standard spelling, in bann, dimounn, kouzinn, is the orthographic reflex of exactly this. It marks an oral vowel followed by /n/, as against the single n of larzan or dipin, which marks a nasal vowel. There is more on that on the alphabet page.

Consonants

There is no /ʃ/ and no /ʒ/ in the inherited vocabulary

This is the single most consequential fact for anyone coming from French, and the point on which most informal descriptions go wrong. French /ʃ/ and /ʒ/ are realised as /s/ and /z/ in their Mauritian reflexes: cheval gives seval, jour gives zour.

/ʒ/ is absent from the language and the 2011 orthography provides no grapheme for it at all. /ʃ/ is marginal and confined to English loanwords, where it is written sh: shoping, kash.

The affricates are recent arrivals

/tʃ/ and /dʒ/ are in the inventory, written ch and j. Old texts suggest both were formerly very rare, and that increasing numbers of non-French words containing them have become established. So the orthography's assignment of ch to /tʃ/ and j to /dʒ/ is not a re-labelling of the French sounds. It is a system built on the premise that the inherited lexicon has /s/ and /z/, and that the affricates enter through other lexical strata, chiefly English and the Indian languages.

Four segments with qualifications

• /t/ and /d/ are dental, as in French, but both palatalise to [tj] and [dj], or lightly affricate to [ts] and [dz], before [i] or [j]. The dental to palatal region of the stop system "appears to be undergoing change", so this is an area in flux rather than a settled contrast.
• /ɲ/ is phonemic but its realisation is unstable. In the long form of gagne, "get", the most usual pronunciation is probably [ganje], but [gaɲe] and [gæɲe] are also current. In the short form gagn it is typically [gaɲ], with [gæ̃j] among the variants.
• /ŋ/ is a full phoneme, written ng, and common in the non-French vocabulary.
• /h/ is marginal. It occurs in some words of English or Indic origin, and in ha as a variant of sa, but its pronunciation is optional throughout, and the ha variant "appears to have only limited geographic distribution".

The letter r is not a French r

In most positions r is today a voiced velar fricative, [ɣ]. Word-finally and before a consonant it is not realised as a consonant at all, showing up instead as a change in vowel quality or as length on the preceding vowel. Note the word "today": this is a change in progress, not a fixed property.

Stress

Mauritian Creole is a stress-timed language, and its stress system is not a syllable-counting rule. Baker and Kriegel describe three kinds of syllable:

1. Unstressable in all circumstances. The tense and aspect markers, subject pronouns, possessive adjectives, the determiners enn and la, the final syllable of the long form of verbs that have both a long and a short form, and most prepositions.
2. Obligatorily stressed. The prefixes re- and de-, the emphasiser mem, the numeral enn meaning "one".
3. Stressable. All nouns, adjectives, verbs and adverbs contain at least one stressable syllable, which may be pronounced without stress next to another stressed word, depending on the rhythm the speaker adopts.

So stress is both lexical and phrasal, and the phrasal part is speaker-variable.

Stress is contrastive, and the orthography cannot write it

This is the most interesting consequence, and it follows directly from the two facts above. Stress placement distinguishes words that the standard spelling renders identically. Using Baker's transcription, where the acute marks the stressed syllable:

• refér, "to recover from an ailment", against réfer, "to make or do again"
• la, the definite determiner, against lá, the locative adverb
• kúd-kúd, "do a lot of sewing", against kud-kúd, "do a bit of sewing from time to time"

The 2011 orthography prescribes no diacritics, so each of these pairs is a set of homographs in standard spelling. Those accented forms above are Baker's phonetic transcription, in which the accents mark stress and intonation contour. They are not Kreol spellings and should not be reproduced as such.

On the often-repeated claim that stress falls on the final or the penultimate syllable: no source states either as a rule. Baker's own transcriptions do put the accent on the final syllable of most polysyllables, so a tendency is visible in his data, but réfer and kúd-kúd contradict it as a generalisation.

An open question worth knowing about

Many Mauritian Creole verbs have a long and a short form, and what conditions the choice is genuinely disputed.

Corne proposed in 1982 that the alternation is prosodically driven. Fabiola Henri argues against a purely prosodic account, and van der Wal and Veenstra agree with her, on the grounds that a following unstressed pronoun does not trigger the long form in the way a stress-based account predicts. Their settled position is that prosody "may indeed be a factor in the alternation… but it can certainly not be the only one". Henri goes further, holding that the alternation "is neither constrained by phonotactics nor phonetic structure of the language in the environment in which the truncation occurs".

The syntactic side of the alternation is covered on the syntax page.

Processes that are actually attested

A short list, because there is a lot of plausible-sounding invention on this subject online. These are the processes described in the literature:

• Denasalisation of historically nasal vowels, in the two environments above. The flagship process.
• Depalatalisation of French /ʃ/ and /ʒ/ to /s/ and /z/ in inherited vocabulary.
• Unrounding of the French front rounded vowels.
• Palatalisation or light affrication of /t/ and /d/ before [i] and [j].
• Article agglutination, which reshaped the word forms of hundreds of nouns. Covered in detail on the French to Kreol page.
• Schwa insertion in non-basilectal varieties, an ongoing change.
• Lenition of /r/ to [ɣ], with loss and compensatory vowel effects word-finally and before consonants.
• Elision in the negator. Pa regularly combines with ena "have", ankor "yet" and the perfect marker, giving forms Baker writes as p'ena, p'ankor and pa'n.
• Phrasal destressing of stressable syllables next to a stressed word.

How the sounds map onto the spelling

The 2011 orthography is close to phonemic but not identical to it. The mismatches are worth knowing.

One grapheme covering more than one sound:

• e covers [e] and [ɛ]; o covers [o] and [ɔ]
• r covers [ɣ], vowel colouring and vowel length
• t and d cover the dental and the palatalised or affricated realisations
• x covers /gz/ and /ks/

Sounds and features with no grapheme at all:

• Schwa, so the dilo variation cannot be written
• [ʒ], which the language does not have
• Stress, which is contrastive, so the pairs above are homographs

Where the same sound is spelled differently depending on the source language. The report is explicit that a word of French origin takes di-, as in media, diab, diaman, while a word from English or an Indian language takes j. The orthography encodes etymology at this one point.

Variation

The evidence here is thinner than you might expect, and most of what is described is change in progress rather than stable regional difference.

• Register. Schwa is absent from basilectal speech and increasingly frequent in other varieties. This is the one properly sourced phonological anchor for the acrolect and basilect distinction. There is no published phonological study of an acrolectal Creole closer to French, so be wary of accounts that assume one.
• Geographic. Only the ha variant of sa is described as geographically limited.
• Change in progress. The stop system between dental and palatal, the rising frequency of /tʃ/, /dʒ/ and /ŋ/ through contact, and the current realisation of /r/.

Where to go next

• The alphabet and spelling system, the 2011 orthography in full
• Kreol: Syntax, how sentences are built
• From French to Kreol Morisien, the sound correspondences and article agglutination
• Is Mauritian Creole French?
• An annotated bibliography of Mauritian Creole linguistics
• All the Morisien guides

Sources

• Philip Baker and Sibylle Kriegel, "Mauritian Creole", Atlas of Pidgin and Creole Language Structures Online, survey 55. The source for the phoneme inventories, the allophony, denasalisation, the stressability system and the stress minimal pairs. Free to read.
• Philip Baker, Kreol: A Description of Mauritian Creole, London: C. Hurst and Co., 1972. vii, 221 pp.
• Philip Baker, "Accounting for contact-induced changes in Mauritian Creole", paper presented at Grammatical hybridization and social conditions, Max Planck Institute for Evolutionary Anthropology, Leipzig, 16 to 18 October 2014. Slides, free. The source for the vowel reduction relative to French and for the rise of the affricates.
• Lortograf Kreol Morisien, Akademi Kreol Morisien, Ministry of Education, April 2011. Free PDF. The source for the grapheme inventory and the di- and j rule.
• Jenneke van der Wal and Tonjes Veenstra, "The long and short of verb alternations in Mauritian Creole and Bantu languages", Folia Linguistica 49(1), 2015, 85 to 116. DOI 10.1515/flin-2015-0003.
• Fabiola Henri, "Morphomic structure in Mauritian Kreol: on change, complexity and creolization", Morphology 31(4), 2021, 447 to 489. DOI 10.1007/s11525-021-09383-9. The sources for the dispute over the long and short forms.
• Chris Corne, Essai de grammaire du créole mauricien, Auckland: Linguistic Society of New Zealand, 1970. 57 pp. Te Reo Monographs.
• Rachel Eva Strandquist, Article Incorporation in Mauritian Creole, MA thesis, University of Victoria, 2005. Free PDF.

Full verified bibliographic details for the works cited here, and for the wider literature, are on the annotated bibliography page.

Rebuilt September 2026. Transcriptions in square brackets and accented forms follow Baker and Kriegel; Kreol words are given in the 2011 standard orthography.