This article proposes that a major drive in the fast evolution of cinema is that film uniquely fits, exploits and expands the potential of a specialized cognitive machinery in the human brain. This is working memory (WM), a limited capacity processing system that temporarily holds and processes on-line and off-line information under attentional control during the planning and execution of a task. A dominant model of WM depicts multiple components, including a central executive, subordinate workspaces for spatio-visual information and for sound and language, and an episodic buffer that binds episodes on the go and is capable of sorting them into long-term memory. The distinct generic attributes of film and their relevance to the subcomponents and operation of WM in the spectator are described. It is proposed that in watching a movie, WM operates in a special mode, dubbed the representation-of-representation (ROR) mode, in which normal motor response to reality is suppressed. It is further proposed that under proper contextual settings and mind set, the central executive of the spectator relinquishes control to the film information, culminating in a transient rewarding dissociative state. The usefulness of the model is discussed in the framework of the newly emerging discipline of neurocinematics. In evolutionary context, the interaction of film and brain is bidirectional. Film in its broadest sense is an extra-corporeal audiovisual space that allows the human brain to perform detailed past and future mental time travel which, unlike WM and human memory in general, has unlimited capacity, variability and endurance. This augments the original phylogenetic advantage that had probably led to the emergence of episodic memory in the first place.