The reciprocals based on 'companion' in the Romance-based Creole languages

International audience Seven out of the seventy-six contact languages treated in the Atlas of Pidgin and Creole LanguageStructures (APiCS, Michaelis & al. 2013) have a special grammaticalized reciprocal markerbased on ‘companion’ (e.g. Cape Verdean Creole of Santiago Es kre kunpanheru. ‚Ils s’aiment’,Lang 2009:180). In this contribution, I will have a closer look at the five Romance- basedCreoles in this group, the three Portuguese-based varieties of Cape Verdean Creole, FrenchGuianese Creole and most importantly Seychelles Creole (French-based). In Seychelles Creole,the grammaticalized use of kanmarad (< Fr. camarade ‘comrade, companion’) as a reciprocalmarker is mentioned in all three available grammars of this language (Bollée 1977; Corne 1977;Choppy 2009). Its evolution and distribution are discussed in Kriegel (2021) with present-daydata from spoken and written corpora and historical data from closely related MauritianCreole. While the companion-reciprocals in Cape Verdean Creole and in French GuianeseCreole have been interpreted in terms of substrate influence respectively from Bambara (Lang2009:180) and from Ewe (Jennings & Pfänder 2018:150), such a substrate explanation forSeychelles Creole is not probable: the main substrates Malagasy and Eastern Bantu languagesdo not have reciprocal constructions with ‘companion’. Malagasy uses the infix -if- incombination with the prefix m- (Jaozandry 2014). As for Eastern Bantu, the dominant patternmarked by –ana is clearly not related to ‘companion’ but goes back to a NP conjunction marker(Maslova 2007:350). The few speakers of West African varieties who arrived in the IndianOcean in the 1730th are supposed to have been speakers of Mandinka, which does not have‘companion’-reciprocal construction either, but a pronoun corresponding to ‘one another’(Creissels & Sambou 2013:212). The grammaticalization of kanmarad in Seychelles Creole willbe considered to be an instance of accelerated functionalization typical of creole languages(Michaelis & Haspelmath 2020). Ultimately, the study’s findings will be discussed in light oftwo complementary hypotheses that try to explain the acceleration of functionalization: theExtra-Transparency Hypothesis (Haspelmath & Michaelis 2017) and the Distinction duringCodification Hypothesis which I suggest for Seychelles Creole. Both are considered to be possiblefactors favoring an ordinary language-internal grammaticalization process.