The Origins of Negation in the Gulf of Guinea Creoles


Tjerk Hagemeijer

(Universidade Autónoma de Lisboa)



It is well-known that the Gulf of Guinea Creoles (GGC) exhibit patterns of discontinuous/final negation, which should therefore be considered an early feature. Whereas it is rather uncontroversial that preverbal negator na has it roots in Ptg. não,  the final marker’s phonetic shape (fa/wa) and its highly specific syntax and pragmatics appear to be unrelated to Portuguese. Discontinuous or final negation only exists in a very limited number of Creole languages (Palenquero, Afrikaans, Berbice Dutch) and is generally assigned to substrate influence (see Schwegler, Den Besten, Kouwenberg, etc.). Moreover, Creole languages with European lexifiers, except for Berbice Dutch, borrowed their negation markers from the lexifier.

            In the case of the GGC, several authors (Boretzky, Ferraz, Stolz) have looked at surrounding African languages that exhibit discontinuous negation, like Ewe (Kwa), Yoruba (Yoruboid), Jaunde and Yaka (Bantu). In none of these languages i) anything similar to fa has been found, ii) these languages cannot be considered relevant substrate languages and iii) no systematic syntactic/pragmatic comparison has ever been provided. Kikongo and Kimbundu, on the other hand, do have discontinuous negation and their presence in the GGC has left imprints, but we believe these imprints to be essentially of lexical nature but not crucial to structural features like negation. Moreover, the main objection to the Bantu hypothesis is Lung’iye (=Principense). Despite all its non-Bantu features, it exhibits final marker fa. Bantu may have had some impact on the negation patterns in Santome, Ngola and Fa d’Ambô, but we believe this to have happened only after the formative period of the proto-GGC. 

            This leaves us with two languages: the presumably oldest and most important substrate at a lexical and structural level, namely Edo (Edoid), and the lexifier, Portuguese. Modern standard Edo only exhibits preverbal negation and no markers whatsoever that follow negation. From the available work (Agheyisi, Melzian, Dunn, etc.), an Edo etymology for fa seems unlikely. If anything, the syntactic distribution of fa reminds the distribution of sentence level emphasis markers in the GGC, rather than, for example, the syntax of polarity items in Portuguese. In the light of the available evidence, we will therefore address the following interrelated hypotheses :


-         fa was initially a negative or affirmative pragmatic marker that specialized for both negative and non-negative environments;

-         negative fa and affirmative emphasizers fa(n)/f(a) (in Santome and Fa d’Ambô respectively) are unrelated;

-         Fa has a Portuguese etymology; candidates to be considered are fava (Schuchardt 1882), falar and fora;

-         Discontinuous/final negation in the GGC is the result of internal development.


Keywords: Gulf of Guinea Creoles; discontinuous negation; syntax; pragmatics; etymology.