[1]
Further, Dresler (this collection) raises the question of whether the frequency of specific dream contents can be regarded as evidence for the importance of its underlying functions. If we consider the function of dreaming more broadly to be that of a training ground for essential and adaptive behaviors, it becomes rather clear that the observed frequency of these behaviors can be viewed as a valid measure of their importance. This, however, is evident only when comparing the contents within the phenomenal level of explanation.
[2]

We should, however, also keep in mind the option that dreaming does not serve any function at all and was not selected for, but is merely epiphenomenal, as suggested, for example, by Flanagan (2001), and implied by the Continuity Hypothesis (CH). This notion should be the null hypothesis against which the proposed functions of dreaming are to be pitted.

[3]
The multifunctionality of dreaming might be possible in different populations, so that in a population that lives in a very threat-filled environment a strong threat simulation system would be selected for, whereas in a population living in more peaceful conditions the psychotherapeutic function and taming of threat simulations dreams would be more likely candidates for selection. However, one and the same population cannot manifest both functions at the same time. Just as in some species of moths, in one environment individuals are selected for towards being white because white provides the best camouflage, while in another environment the color of individuals in the same moth species is selected for towards being dark gray or black, because in that environment all the white individuals are too easily detected by predators.