A microanalysis of fourteen dream reports was conducted with the aim of understanding which features of these reports rendered them more or less narrative. The result should help us to compare the relative narrative quality of samples of dream reports. The reports described the last dream of the night and they were collected immediately upon spontaneous morning awakening in a sleep laboratory. Half of the dreams reported in our sample were produced in Rapid eye movement (REM) sleep and the other half in Stage 2 sleep. The microanalysis dealt with the relationship between each pair of successive temporalunits and with the presence of other elements of the dream content likely to influence the narrative character of a dream. The more important findings were the following. First, coherence is the rule in the relationship between pairs of successive units of the reported dream sequences. Second, two categories of relationship, Psychologically causal links and Plausiblelinks appear to be the main building blocks of the sequences of dream events. Third, most dreams comprised one or several complications, that is, events creating change and often tension. Fourth, the combination of a continuity score and a complication score permitted us to rank the reports according to their narrative nature or quality, whereas emotions and story-like features were not good candidates to that purpose. Fifth, no differences were observed between REM-sleep and Stage 2 reports of similar length in their sequential regulation and other features. This result shows that the process regulating thesequences of dream events is not dependent on the sleep stage, a strong argument in favor of the idea of a single generatorwhich produces dreams across the sleep stages. More generally, most sequences of dream events did not correspond to sequences of recent or older experiences stored in memory. The sequential organization of dream events is regulated during dreaming and reveals an important dreaming production process.
