Cohesion can be described as the explicit indication of textual coherence by means of linguistic devices. One of these devices is the use of different tenses: tenses help to structuralize the discourse (text) by locating individual States of Affairs in time and relating them to each other on various levels. These levels can vary between the individual sentence (micro level) and the wider context of the discourse. A well-known example of this structuralizing use of tenses is the alternation of imperfects and aorists in Greek narrative texts: the imperfects create a frame (have a backgrounding function) for the aorists, which, in turn, mark the successive steps in the narrative (see e.g. Rijksbaron (2002: 11-14)). Unfortunately (and surprisingly) the attention paid to the cohesive force of Greek tenses is virtually limited to the narrative tenses (viz. aorist, imperfect and historical present). The perfect has not been analyzed from this specific viewpoint and generally speaking is treated quite stepmotherly in the literature. In my contribution to the 6th International Colloquium on Ancient Greek Linguistics I will focus on the use of the perfect in various (non-narrative) texts and argue that the commonly accepted description of the Greek perfect needs refinements. Moreover, I will attempt to show that the perfect functions as a cohesive device on various discourse levels. The communis opinio about the semantic value of the perfect can be resumed as follows: the perfect stem signifies both that a state of affairs is completed and that as a result a state exists (stative-confective value); the primary perfect indicative (commonly: perfect) locates the state at the moment of utterance (the present ) (Rijksbaron (2002: 1;4); cf. e.g. Ruijgh (1971: 229-30); Rijksbaron (1984: 403-4)). This definition is however rather unsatisfactory for a large number of perfects. In many cases, the idea of a resulting state is inapplicable, or forced at the very least.