RANDOM COGNITIVE ACTIVATION IN DREAMING DOES NOT REQUIRE A CARTESIAN THEATER

Recent theories of dreaming have suggested that dream content is deter mined by the random activation of elements within a mnemonic network. These theories then suggest that the cortex is somehow able to make se nse out of these random patterns of activation and that during this pr ocess the brain constructs as coherent and plausible a dream narrative as possible. Such an approach assumes that there is indeed a place in the brain where all the information ''comes together'' into a unified experience and from where the dream events may be ''witnessed.'' Denn ett (1991) has named such a place the Cartesian Theater and has demons trated that for the case of waking consciousness such a notion is fund amentally misguided. Dennett's arguments can be broadened beyond wakin g consciousness to include the sleep state, and it is proposed here th at there is no such place where the random activations are woven into a coherent story and that there is no Cartesian Theater in the dreamin g brain. Instead, it is proposed that dream content is synonymous with which elements of the connectionist-style mnemonic network are most h ighly activated and that the threshold for a given element is determin ed (at least in part) by the recency or saliency of the stimuli repres ented by that element.