Abstract

In a series of previous experiments (Mazzoni and Loftus, 1996) we showed that dream words can be falsely recognized. In addition, the subjective judgment was different for false recognitions of own-dream words (dream-self) than for false recognitions of words taken from dreams of other subjects (dream-other). Using the Remember/ Know judgment (Tulving, 1985) dream-self false recognitions were mostly judged Remember, whereas dream-other false recognitions were judged mainly Know. In this study we constrasted the hypothesis that the Remember/Know difference is specific to dreams, with the hypothesis that the effect extends to most self-generated materials (Slamecke and Graf, 1978). Two experiments that used dream words and story words contrasted the two hypotheses, and showed that dream words produced a higher rate of false recognitions. The quality of such memory errors however did not differ in the two conditions, with self-generated material having the same judgments independently of it being part of a dream or part of a story. Implications of these results are discussed in terms of Johnson's (Johnson, Kahan, Raye, 1984), Tulving's (1985) and Jacoby's (Jacoby and Yonelinas, 1995) proposals on the mechanisms that can create memory errors.