Negation in Musgum

This paper studies the negation construction in musgum language.We collect the musgum data on negation from native users and analyse them in terms of Kayne’s (1994) antisymmetry and Rizzi’s (1997) split CP approaches. We identify the free negation element (á:à) and several negation markers (kài, kirkài, kài tiŋ and kirkài tiŋ) that close independent and complex clauses. In complex structures with completive and relative clauses, the main clause cannot contain a negation marker. In complex structure with adverbial clause, negation marker can be present in main and adverbial clauses. We discover that Negation Phrase is the highest projection, higher than Force Phrase, rejecting the split‐CP projections order of Rizzi (1997). When the negation head is generated , Inflexion Phrase is subject to heavy pied‐piping. It occupies the specifier of Negation Phrase.

projection IntP (Interrogative Phrase). The specifier position of which can host specific interrogative operators such as perché "why" in main and embedded clauses or interrogative particles such as se "if" or "whether" (in embedded clauses in Italian).
Biloa (2013), analyzing tuki (Bantu language spoken in Cameroon), states that questions markers do not occupy the Force Phrase (Force P). They rather occupy the head of Focus Phrase (FocP). This fact is different from the Rizzi's (2013) viewpoint.
Ondoua Engon (2013) tackles the negation phrase in bulu (Bantu language spoken in Cameroon). He shows that the Negation Phrase is higher than Force Phrase. As for him, the Rizzi's viewpoint does not accommodate the bulu data. He notices that the projections order can change from a language to another.

Methodology
In generative grammar, the methodology aims at description and explanation. We used two stages. The first stage was to collect the musgum data on negation construction. For doing this, we used interrogation. We questioned the native users of musgum language and we take their answers.
The second stage concerned the data analyzis. We presented them, described them and commented them. According to Laenzlinger (2002), we analyse the empirical data by describing them so that we obtain some generalizations. In fact, we carry out a structural description that can permit us to have some relevance. Then, we propose hypothesis by considering our generalizations. Finally, we check the hypothesis by making compatible theory and data.

Negation markers
Musgum language has several negation markers.
"Kirkai" is used for expressing negation of process that was ongoing but ceased at the moment of declaration. Let us observe the following data: Woman poss. SM loves obj. neg.
"His wife does not love him more" "kai tiŋ" is expressing a process that can never happen or take place. We have the following examples: 3SG laughs neg. never "He never laughs" b. Zigla a píyí-tí am-ni kài tiŋ.
Zigla SM seeks ObjM mother his neg never "Zigla never seeks his mother" "kirkai tiŋ" is more complex. It is used for asserting a process that was ongoing before, but can never more continue. We have the following data: a. Musa a suma hilif kirkai tiŋ.
3SG does work neg never "He never works more" "duk kai" expresses a process that never happens. Let us take the following data: a. à mùdà dùk kài 3SG spoke one time neg "He never spoke onetime" b. fáŋ á slá dúk kài rain SM falled one time neg "It never rained onetime"

Negation in complex structure
A complex structure has less than two verbs. It contains a main clause and one or several subordinate clauses. The subordinate clause can be completive, relative or circumstantial.

Structure with completive clause
Let us consider the following complex structures: 3SG thinks neg that death SM prog. waits "He does not think that the death is waiting for" The first structure (8a) is the basic one. It has a main clause "á híná" (he thinks) and a completive clause "ɓó mírí á ná báyá" (that the death is waiting for). In the second structure (8b), negation morpheme "kai" (not) is in the completive clause domain and it is acceptable. But in the third structure (8c), the negative morpheme "kài" (not) is in the main clause "á híná kài" (he does not think). This structure is grammatically incorrect.

Structure with relative clause
Let us observe the following data: 2SG gets married man that 2SG love ObjM neg "You get married à man that you do not love" The first structure is the basic one (9a). It contains a main clause "kí tíká wús" (you get married) and a relative clause "nà kú fúryáŋ" (that you love). In the second sentence (9b), the negative morpheme "kài" (not) is in the main clause. The structure (9c) becomes grammatically incorrect. But in the third sentence, the negative morpheme "kài" (not) is in the relative clause domain. The structure is grammatically correct.
Let us notice that in the first sentence, the relative clause is not restrictive. In this case, a negative morpheme cannot concern the main clause. The fact is not the same thing: we have affair to a structure that the relative clause is restrictive. We have the following data: Woman that SM came SM prog. cries neg "The woman who came is not crying" In the basic structure (10a), we have a main clause "míníy tí ná túwá" (the woman is crying" and a restrictive relative clause "nà tì hàlàsì". In the following sentence (10b), the negation morpheme "kài" (not) is in the relative clause domain. The structure is grammatically correct. In the other following structure (10c), the negation morpheme is in the main clause domain. It is also correct.

Structure with adverbial clause
Let us consider the following structures: 3SG runs neg. because 3pl gave him neg.
"He does not run because they did not give him" In the first structure (11a), we have main clause that contains the negation morpheme "á láwá kài" (he does not run), and an adverbial clause "bríyɓó à kàyà sù" (because he was tired). This sentence is syntactically and semantically correct.
In the second structure (11b), the negation morpheme is whereas in the adverbial clause domain "bríyɓo à kàyà kài" "because he was not tired). The structure is also syntactically and semantically correct.
The third sentence (11c) has respectively main and adverbial clauses: "á láwá kài" (he does not run) and "bríyɓó è sìŋ kài" (because they did not give him). Both main and adverbial clauses have negation morpheme "kài" (not). The sentence is syntactically and semantically correct.

The structure of Negation Phrase (NegP)
When we observe negation marker in independent sentence, it is at the end: (12) Margaza à hàlà kài.
"Margaza did not go" The morpheme kài (not) represents the negation phrase. In the surface structure, it concludes the structure; but in the deep structure, it occupies a high position. We can have the following tree representation: This fact can be accounted for Chomsky's (1973Chomsky's ( , 1993 Extension Condition: "when a head is merged, movement into its specifier is obligatory". We can also call to mind Kayne's (1994) Linear Correspondence Axiom (LCA): AgrP goes to Spec-NegP because the head should be preceded by its specifier and followed by its complement. The pied-piping of IP into the specifier position of NegP can also be accounted for by the EPP (Roberts, 2001).
We noticed previously that the negation morpheme in main clause of structure with completive clause is not acceptable. But it can be at the end of complex structure with completive clause. Let us re-consider (8b): 3SG thinks that death SM prog. waits neg "He thinks that the death is not waiting for" For such structure, we can have the following tree representation: The tree representation shows that NegP is the highest projection. Does it mean that the Negation Phrase is higher than the Force Phrase? For understanding the fact, let us consider the following structure: (16) ɓó mírí á ná báyá kài.
that death SM prog. waits neg "[…] that the death is not waiting for" We indicate that at the surface stage of derivation, "ɓó" (that) begins the structure and negation morpheme "kài" closes it. But in deep stage, the fact is not the same. Let us observe the following tree representation: Generally, several morphemes mark negation in musgum language. Except the first free element (á:à), the others occupy the end of the sentence. In complex structure with relative or completive clause, the negation morpheme cannot occupy the main clause domain. But in complex structure with adverbial clause, both main and adverbial clauses can contain negation markers. Those markers conclude each domain. The fact shows that Negation Phrase is higher than Force Phrase considered by Rizzi (1997) as the first projection of his Split-CP diagram.

Conclusion
We identify several negation markers in musgum language: á:à, kài, kirkài, kài tiŋ and kirkài tiŋ. Except the first free element (á:à), the others occupy the end of the sentence in phonologic form. In complex structure with relative or completive clause, the negation marker cannot occupy the main clause domain and remains at the end of the construction. But in complex structure with adverbial clause, both main and adverbial clauses can contain negation markers. Each clause is concluded by a negation element. According to Kayne's (1994) Antisymmetry, they are generated in a high position, precisely at the left side of Inflexion Phrase (IP). The fact shows that Negation Phrase is higher than Force Phrase considered by Rizzi (1997) as the first projection.