Much research in the past has been devoted to the use of reported speech and thought in fictional writing (Banfield, 1973, 1982; Volosinov, 1973; Leech and Short, 1981; Sternberg, 1982; Fludernik, 1993 – to mention only a few) and in everyday oral story-telling (Tannen, 1989; Mayes, 1990; Holt, 1996). In oral story-telling prosodic dimensions of speech have been shown to be instrumental in marking quotations off from surrounding talk (Klewitz and Couper-Kuhlen, 1999), in constructing enacted scenarios (Yule, 1995; Couper-Kuhlen, 1999) and in conveying multi-voicedness (Maynard, 1996; Günthner, 1999; Holt, 1999). Yet the type of reported speech and thought examined in these studies has as a rule been that found in narration, where a sequence of events is recounted in chronological order and participants in these events are animated as story-world figures. Little or no attention has been paid to isolated quotations in conversation which are not embedded in a story framework nor are part of any larger narrative context. It is this kind of speech and thought-reporting which the present study sets out to investigate. It will be argued that reported speech in non-narrative contexts is different in design from that in narrative contexts and that it is frequently incorporated into assessments and accounts as a means of heightening evidentiality. In contrast to many forms of narrative reported speech it will be shown that the boundaries between the ‘there-and-then’ of non-narrative reported speech and the ‘here-and-now’ are fuzzy.