In most Bantu languages the subject marker is a mandatory element affixed to the verbal base. In Batoufam, we find a subject marker suffixed to the verbal root only in some specific contexts. The marker may take either a ts- or a zh- form, which is similar to the morphology of some noun classifiers. By being attached to the verb, the marker would normally be expected to play a grammatical role in the sentence. Our examination of the different contexts of occurrence shows, however, that the absence of the said-marker has no impact on the grammaticality of the sentence. Though the grammar requires an agreement of class in number and person between the subject and the subject marker, in the cases examined what we find instead is just an agreement of person.

In the first example below the sentence is perfectly grammatical despite the absence of subject marker.

1. s a Ä m -E m - s h i X u s ö ö

"When is my fufu going to cool down?"

2. s a Ä m -E m - s h i X u -z h -« s ö ö

Examples 2. above shows how, when it occurs, the affix of class on the subject marker z h -« does not agree with the grammatical subject of classes 4 (m- is the affix marker of agreement for class 4). The marker cannot be translated and its meanings vary according to the discourse context. In examples 2. above, the use of the marker indicates that the speaker is somehow unhappy with the situation described by the proposition.

Although the marker can be used to describe an array of attitudes (like admiration, envy, criticism, surprise, unhappiness), its unique function is to express the speaker’s attitude towards the grammatical subject of its embedded clause, in relation to the event or events being described. This morpheme is not a subject marker in the usual sense. The author in a discourse might or might not be an argument of the verb, the marker, therefore, cannot be analyzed exclusively in terms of agreement between the verb and its arguments. This would explain why we do not have the expected agreement of class between the subject and the marker. As the expression of the relationship between an author and his/her discourse, this particle requires an analysis beyond the boundaries of a clause or a sentence. Consequently, the approach that would be the most suitable, and that was used for our inquiry is discourse grammar.