rominence. Williams et al., when comparing dream reports with waking fantasies, showed that discontinuity is the most state specific class of bizarreness, being 6 times more frequent in dreams than in fantasies (1992). Our results failed to support his finding, with incongruity leading as the most state specific class of bizarreness.At the class level, bizarre transformations of objects and characters appear to be controlled by associational constraints that require the transformed item to normally remain within the same class after the transformation. He also found that no transformations of inanimate objects into characters or vice versa were observed (Rittenhouse, Stickgold & Hobson, 1994). Our findings support this study, however our sample indicated that only 12.5% of participants recalled transformations in their reports. This sample is too small to show an accurately high significance.Our inability to confirm our hypothesis may have been due to such methodological errors as 80 participants was not a large enough sample to gain accuracy. It is recognised that reports may have been edited in order to prevent embarrassment, however this is not regarded as likely to confound our analysis. Another problem with using home-based reports is the lack of controlled conditions. Each subject experienced different settings and report techniques. More importantly we cannot ensure that participants recorded their experiences immediately after they occurred, this might have resulted in state dependent amnesia.In summary, this study failed to support Hobsons activation-synthesis hypothesis that there would be greater bizarreness in dreams than in daydreams or waking narratives. The main psychological finding is that, contrary to the activation-synthesis hypothesis; dreams did not have more bizarre features per sentence than daydreams. The one category in which dreams exhibited greater bizarreness was the category for incongruities. Transformation...
