The “problem” of dreaming in NREM sleep continues to challenge models that propose a causal relationship between REM mechanisms and the psychological features of dreaming. I suggest that, ultimately, efforts to identify correspondences among multiple levels of analysis will be more productive for dream theory than attempts to reduce dreaming to any one level of analysis.