This paper adopts a theoretical perspective on Discourse Markers (DMs) which is currently being worked out by Fraser and Kalokerinos (in preparation). According to this approach, DMs are linguistic expressions, which serve the integration of meaning units (MU1) to upper level meaning units (MU2). Typically, DMs integrate two MU1 into one MU2. MU1 consist of at least a complete proposition. Prior to the process of integration, at least one MU1 must have an illocutionary force (i.e. be endowed with a discourse modality) independently of the other MU1. If this is not the case, then integration takes place at the sentential level, which precedes the level of discourse. During the process of integration in discourse, the modality of one or of both MU1 may get specified. This is the part of the integrative process which Fraser and Kalokerinos name ‘interpretation by integration’.

According to the aforementioned authors, one main category of DMs is ‘implicatives’. This category consists of two main classes, namely, ‘causals’ and ‘inferentials’. The members of both classes serve to establish the discourse relation of justification. Discourse integration though is achieved with different, and indeed complementary, means of discourse syntax. As a result, the members of the two classes serve the construction of different discourse perpsectives. That is to say, they bring about different strategies of argumentation.

The present paper focuses on the specific properties of some DMs belonging to the two aforementioned classes of implicatives from Modern Greek, English and French. It explores the ingredients of their procedural meaning, which allow speakers using them to reach aspects of the non-explicit meaning of discourse. These functions range from the uncovering of presuppositions to metacommunication.. Bringing into focuswhat has been implicit information may serve various argumentative purposes in monologue and dialogue. It helps to a construction of different images of the preceding discourse and highly constrains its possible continuations. It thus appears that, although DMs are optional in discourse, their presence does more than simply constrain interpretation (a claim of relevance theorists) or speeding up processing (a claim of psycholinguistically oriented coherence theorists).

The DMs under consideration are Modern Greek jati, afu, oste and tote, English because, for, since, and then, and French car, puisque, and alors.