Culler starts by saying that among the different scholars in narratology, one thing they all agree on is the distinction between 'story' (a sequence of actions or events independent of their manifestation in discourse) and 'discourse' (the discursive presentation or narration of events). Traditionally, the analyst of narratives must, after having distinguished between narratives and non-narratives, make the assumption that the events are independent from discourse and have a chronological, true order. This is what Culler problematizes in the rest of the text.
Using some examples (King Oedipus, Daniel Deronda, and the writing of Freud), Culler argues that these narratives have impact because of a double reading, using two 'incompatible logics'. The suggestion is made that, perhaps, there is no real order in which events take place, only the order in which they are experienced in discourse.
Culler mentions the work of sociolinguist Labov on 'natural narratives'. Labov studied the stories people tell naturally by assuming that there were one the one hand, events, and on the other, a way in which these events were told, resulting in two classes of clauses. Culler shows however that the clauses cannot be sorted neatly into these classes, since it is not clear whether a clause is reporting an event (narrative clause) or is commenting on the events (evaluative clause).
At the end, Culler sums up his argument: in one perspective, the story (a sequence of abstract events) give rise to the discourse (the presentation of those events). In another perspective, the abstract events are a product of the discourse, since it is only the discourse that is directly available to us. Culler argues that these two perspective are equally important in order to account for the impact of narratives, but that they cannot be combined. This puts in question the possibility of a 'science' of narrative. One must be willing to shift between the two perspectives, between story and discourse.