A neodissociative model of mind is better equipped than a social-psychological model to deal with the complexities of hypnosis, and of human behavior generally. It recognizes, as Coe's (1992) model does not, that behavior can be more automatically activated than strategically enacted. In particular, Coe's emphasis on human behavior as purposeful and goal directed does not distinguish between goal-directed behavior that serves a purpose, and goal-directed behavior that is performed on purpose. It is this distinction that permits goal-directed behavior to be dissociated from a person's conscious plans and intentions. In addition to offering a critique of Coe's "limited process" view of hypnosis, 4 main points are made in the interest of developing a slightly modified, neodissociation view of hypnosis. First, it is argued that goal-directed fantasies are more limited in their ability to mediate hypnotic responding than is commonly appreciated; as well, they do not seem to account for the nonvolitional quality of hypnotic responding. Second, it is argued that hypnotic ability is not unidimensional, with compliance and social influence more apt to account for the low than for the high hypnotizable's responsiveness to suggestion. Third, compared to low hypnotizables, the hypnotic responsiveness of high hypnotizables seems more likely to result from dissociated control. In other words, for high hypnotizables, hypnotic suggestions may often directly activate subsystems of cognitive control. Consequently, the need for executive initiative and effort to produce hypnotically suggested behavior is minimized, and such responses are therefore experienced as nonvolitional. Fourth and finally, while goal-directed fantasies typically accompany hypnotically suggested responses, they are in many cases more a marker of dissociated control than a mediator of suggested effects.