While studying dead languages, especially ancient ones, we necessarily have to rely on a number of written documents, preserved by a happy coincidence, from which we try to ‘distil’ the rules of grammar, mostly in an inductive way, frequently by simply grouping similar examples together. These are bottom-up procedures taken to their extreme and in this situation the top–down approach as advocated by Functional Discourse Grammar can offer refreshingly new and unexpected insights. It allows us to see the actual expressions found in texts as ‘written discourse’, as the Speaker’s ‘encoded intentions and conceptualizations’. There are no one-to-one correspondences, however, between the categories of discourse and their encoding by means of specific grammatical rules.

In this paper I would like to discuss correlations between certain textual and grammatical categories as observed in Bohairic-Coptic narrative texts. To which extent are these correlations systematic? How can they be related to the communicative situation in which the texts in question were intended to function? Are they instrumental in attaining the Speaker’s communicative goals? The paper will be divided into three sections.

1. “Sentencing the sentence” (Kwee 2000)
This section will be devoted to the segmentation of narrative texts into paragraphs and parts of paragraphs (setting/introduction, complication, peak, resolution) and to the correspon-dences between these segments and the Moves and Discourse Acts of FDG. Particular attention will be paid to the function of ‘complex beginnings’ in the sense of Smits (2002).

2.  Hidden voices
FDG incorporates the participants in the communicative situation, the Speaker and the Addressee, into the model. While the identity of these interactants is quite obvious in an actual conversation, this is not so in the case of written discourse. It is generally acknowledged, for example, that in narrative texts the Speaker is not the author, but a textual construct called the narrator. Besides, a narrative text can contain embedded narratives told by the story’s characters, who then function as secondary narrators. This section will present, first, the linguistic differentiation between the contributions of the different Speakers. Secondly, it will discuss non-narrative Moves and Acts, such as orientations, comments and descriptions.

3. The hero, the villain and the mob
As it appears, the form and the position of a grammatical subject in Coptic narratives correlate to a great extent with the role which the character referred to by this subject plays in the story. We can distinguish the following types of participants: the hero, the opponent, the heavenly helper, secondary participants, and props. Each of these participants may be introduced into the story by means of a different construction and will be resumed differently afterwards. As such concepts as the hero, the opponent or the heavenly helper are by no means linguistic categories, the different treatment of the respective nominal expressions should be explained by the different degrees of topicality of their referents. This requires a more detailed reflection on Topic marking in Coptic.