According to Revonsuo, dreams are the output of a evoked "threat simulation
mechanism." The author marshals a diverse and comprehensive array of empir
ical and theoretical support for this hypothesis. We propose that the hypot
hesized threat simulation mechanism might be more domain-specific in design
than the author implies. To illustrate, we discuss the possible sex-differ
entiated design of the hypothesized threat simulation mechanism.