The Syntax of Mauritian Creole

This is an overview of Mauritian Creole syntax with sources attached, aimed at linguists and at learners who want the real structure rather than a phrasebook. Every example is referenced. Where the literature disagrees, both positions are named, and where the sources do not settle a question, the page says so instead of filling the gap.

Word order

The neutral order is subject, verb, object. Baker and Kriegel code it as SVO at 100% frequency. What is interesting is the departures.

Focus is morphological, not positional. It is normally achieved by suffixing mem to the item emphasised. Objects and independent pronouns can be fronted, but subject noun phrases cannot. That asymmetry has no French counterpart.

Subjects can be absent. Null subjects occur, as in si met mwa dan lot klima mo pa pou kapav, "if they put me in another climate I will not be able to cope".

Objects can appear clause-initially, and what that construction actually is happens to be the liveliest dispute in the field. See the section on the passive below.

Topic and comment

Diana Guillemin has argued that Mauritian is shifting towards topic prominence, and her data are worth seeing whatever one makes of the analysis.

Guillemin treats se, from French c'est, as introducing a topic and comment structure with a zero copula between them, rather than a true cleft, and notes that se is characteristic of acrolectal Mauritian. The clause-final sa in the second example she analyses as predicate fronting with sa left behind. The data are solid; the analysis is hers and is contested.

The noun phrase

Nouns are invariable. A bare noun may be read as singular or plural according to context, unless the indefinite enn, a numeral, or the preposed pluraliser bann is present to disambiguate.

la follows the noun, and Baker and Kriegel warn that it "is frequently described, somewhat misleadingly, as the definite article". It marks specificity rather than definiteness, and the discontinuous sa … la pattern is by far the more common. Bare nouns can carry definite readings too: commoner in Baissac's 1888 folktales, but by no means rare now.

Adjectives fall into two positional classes, and reduplication does opposite things in each. Preposed, reduplication is augmentative; postposed, it is attenuative.

Possession has two orders, and they combine recursively.

Pronouns

There is no case marking, and one form serves subject and object function. Two things the table hides.

The dependent and independent distinction is a segmental one only in the first and second person singular. Elsewhere the two columns are the same string, and the sources distinguish them by a stress diacritic alone. Since the 2011 orthography writes no stress, the contrast is invisible in writing for five of the seven cells.

Second and third person plural are syncretic in zot, which is a real ambiguity rather than a tidy paradigm.

On to and ou: both are listed for Mauritian, and there is a separate page on when each is expected. Worth noting that Baker and Kriegel's remark on this is comparative, about French creoles collectively, and does not settle how alive the distinction is in current Mauritian usage.

The verb group

Verbs do not inflect for person, number or tense. Tense, mood and aspect are carried by free preverbal morphemes. These are markers, not auxiliaries: they are not verbs and they do not inflect.

The perfect is one morpheme with three shapes, not three markers: finn, reduced inn, and enclitic 'nn as in Mo'nn danse, "I danced". Note that it must be written with the doubled n, because single fin spells a nasal vowel and is a different word.

The markers stack in a fixed order. Attested combinations include ti'nn for the past before past, ti pe for the past progressive, ti ava pe for a counterfactual with past reference, and ti a and ti pou for the past conditional, indefinite and definite respectively. As far as the attested data go the order is: negator, then ti, then finn or the future markers, then pe, then the verb. ti precedes everything; pe sits closest to the verb.

The long and short verb forms

This is the signature feature of Mauritian verbal syntax, it has no French parallel, and it is the thing most short descriptions leave out. About 70% of verbs alternate between a long and a short form: manze and manz, danse and dans.

The condition, in Hassamal and Abeillé's formulation, is that the verb takes the short form when followed by a non-clausal complement and not focalised. Baker and Kriegel put it as short form "when immediately followed by a direct or indirect object", which is narrower, and the data below favour the former.

The clause template that falls out of this: subject, then TMA markers, then any preverbal adverb, then the verb, then complements, then postverbal adverbs. Direct complements precede postverbal degree adverbs; oblique complements may permute with them.

The alternation is not phonological. Fabiola Henri is explicit that it "is neither constrained by phonotactics nor phonetic structure of the language in the environment in which the truncation occurs", and that it "cannot strictly be reduced to phonological or morphosyntactic principles". It is not truncation, not elision, and not a matter of fast against careful speech. Any description that gives manz and manze as free variants is wrong.

The copula

There is no copula in equative, attributive or locative predicates. A form ete appears when the predicate is fronted or questioned, that is, when the slot after the subject would otherwise be empty.

The first two are the clearest illustration available: ete in the question, nothing at all in the answer.

Negation

pa follows the subject but precedes all the verbal markers. This is the point worth getting right: pa sits outside the whole marker stack rather than next to the verb.

French needs two elements, ne and pas, around the verb. Mauritian needs one. pa also fuses with following items, giving pena from pa and ena, "to lack, not to have".

Mauritian is a strict negative concord language, on Fabiola Henri's analysis. Sentential negation is canonically preverbal but appears postverbally with neg-raising verbs, and only preverbal negation participates in concord. Among the negative concord items zame, "never", is idiosyncratic in being able to appear preverbally without pa, where the others require it. This is the current account rather than a settled consensus, and the full inventory of concord items is not something I can give from the sources I could reach.

Questions

Polar questions are made by intonation alone. Any statement becomes a question by switching the intonation on the final syllable from falling to rising. There is no interrogative particle and no inversion. A tag, non?, is also available.

Content questions front the question word. A noun phrase inside the predicate is replaced by ki and moved to the front of the sentence. So Mauritian is a wh-fronting language, not a wh-in-situ one, which is worth stating explicitly because French allows both and learners transfer the French pattern.

And as noted above, that fronting forces the long verb form.

Relative, complement and adverbial clauses

The relativiser is ki. It may be omitted in many environments, but not out of a noun phrase forming part of the predicate.

Look at where la sits in the first example. It closes the whole relative clause, not the head noun. That is the best single argument that la is a clause-level specificity marker rather than an article.

Object complement clauses are either zero-marked or introduced by ki, from French que, including with verbs of speaking and of knowing. Adverbial clauses take a range of subordinators: kan "when", si "if", parski and akoz "because", apre "after", avan "before".

Is there a passive? A genuine dispute

Baker and Kriegel describe a morphologically marked passive using gagn: enn lisien inn gagn bate, "a dog has been beaten".

Anand Syea argues in a 2024 paper that this is not the right picture, and that passive-type constructions in Mauritian are apparent rather than real passives. On his account the displaced theme object in an object-initial sentence sits in topic position, not subject position. His evidence is distributional: the fronted object has to be specific and inanimate, and gagn plus verb is lexically restricted.

Likewise gagn bate, "got beaten", and gagn kriye, "got shouted at", are fine, while a form meaning "got killed" is not. The safe description is that Mauritian has a lexically restricted gagn plus verb construction with passive meaning, alongside object-initial sentences whose analysis is disputed.

Related voice constructions: reflexive li ti kapav enform li mem, "he was able to inform himself", and reciprocal saken ed so kamarad, "they help each other".

What is still open

• The passive, as above. Baker and Kriegel against Syea, not reconcilable as stated.
• What conditions the long and short forms. Corne proposed a prosodic account in 1982. Henri rejects it, and van der Wal and Veenstra concur, arguing that a following unstressed pronoun does not trigger the long form as a stress account predicts. Their position is that prosody may be a factor but cannot be the only one. Henri goes further and treats the alternation as morphomic.
• Whether Mauritian is topic-prominent. Guillemin argues for a typological shift from subject prominence to topic prominence, reassigning la, the copula and the passive together. That is a substantive alternative to the conventional description rather than a rephrasing of it.
• Causatives and verb chaining. Neither is covered by the freely available descriptions, and rather than reconstruct a pattern from intuition this page leaves them out. The absence of Mauritian from the serial verb literature is itself mildly interesting: that literature looks at the Caribbean.

Three things a learner-facing description would miss

1. The long and short verb alternation is the grammar's signature and it is syntactically conditioned. Roughly seven verbs in ten participate. Any account that presents the two forms as interchangeable has missed the central fact about Mauritian verbs.
2. Stress is lexically contrastive and the orthography does not write it. la the determiner and la the locative adverb are distinguished by stress alone, as are the two verbs both written refer. In standard spelling these are homographs. There is more on the phonology page.
3. Focus is morphological, and subject noun phrases cannot be fronted. Focus comes from suffixing mem; objects and independent pronouns front, subjects do not.

Where to go next

• The sounds of Mauritian Creole, phonology and pronunciation
• The alphabet and spelling system
• Is Mauritian Creole French?
• From French to Kreol Morisien
• An annotated bibliography of Mauritian Creole linguistics
• All the Morisyen guides

Sources

Free to read:

• Philip Baker and Sibylle Kriegel, "Mauritian Creole", in Atlas of Pidgin and Creole Language Structures Online, Max Planck Institute for Evolutionary Anthropology, 2013. Survey 55. The backbone of this page.
• Shrita Hassamal and Anne Abeillé, "Degree adverbs in Mauritian", Proceedings of the 21st International Conference on Head-Driven Phrase Structure Grammar, CSLI, 2014, 259 to 279. Free PDF. The precise statement of the long and short form condition.
• Anand Syea, "Absence of syntactic passive in creoles: evidence from French-based Mauritian Creole", Canadian Journal of Linguistics 69(2), 2024, 174 to 202. DOI 10.1017/cnj.2024.13. Free to read.
• Lortograf Kreol Morisien, Akademi Kreol Morisien, Ministry of Education, 2011. Free PDF.

In print or behind a paywall:

• Anand Syea, The Syntax of Mauritian Creole, Bloomsbury Studies in Theoretical Linguistics, Bloomsbury Academic, 2013. The standing monograph, and the place to settle most of what this page leaves open.
• Anand Syea, French Creoles: A Comprehensive and Comparative Grammar, Routledge, 2017.
• Diana Guillemin, The Syntax and Semantics of a Determiner System: a case study of Mauritian Creole, Creole Language Library 38, Amsterdam: John Benjamins, 2011. xviii, 310 pp. DOI 10.1075/cll.38. The monograph-length treatment of the determiner system.
• Diana Guillemin, "Evidence of topic-prominence in Mauritian Creole: a typological shift from French to Creole", in Alleesaib and Lefort (eds), New Perspectives on Mauritian Creole and Reunion Creole, Contact Language Library 61, Amsterdam: John Benjamins, 2025, 49 to 95.
• Fabiola Henri, "A lexicalist account of negation and negative concord in Mauritian", in Déprez and Henri (eds), Negation and Negative Concord: the view from Creoles, Contact Language Library 55, John Benjamins, 2018, 69 to 100.
• 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.
• Daniella Police-Michel, Arnaud Carpooran and Guilhem Florigny, Gramer Kreol Morisien, Volim I, Akademi Kreol Morisien, 2012.
• Arnaud Carpooran, Diksioner Morisien, 3rd edition, 2019.
• Philip Baker, Kreol: A Description of Mauritian Creole, London: C. Hurst and Co., 1972. vii, 221 pp.
• Chris Corne, Essai de grammaire du créole mauricien, Auckland: Linguistic Society of New Zealand, 1970. 57 pp. Te Reo Monographs.

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

Rebuilt September 2026. Examples are respelled into the 2011 orthography and so may differ from the form printed in the source cited.