A brain charged with guiding its body through a complex and lively world from a position of solitary confinement inside its opaque skull faces a set of functional problems whose solution may account for the existence and nature of consciousness. An analysis of the more general and basic of these problems, sensory as well as motor, suggests the utility of implementing a high-level mutual interface between sensory target selection, motor action selection, and motivational ranking of needs at a late stage in the run-up to the brain’s decision about the very next action to take (typically a gaze movement). The three selection processes are subject to a number of mutual dependencies such that a regimen of constraint satisfaction among them would yield savings in behavioral resource expenditure. The logistics of implementing such a regimen can be simplified by casting the interface in a particular nested, analog format. It would host a real-time synthetic summary of the rest of the brain’s interpretive labors, reflecting best estimates of the veridical current state of world, of body, and of needs for purposes of real-time decision making. A number of design principles for the neural mechanism realizing such a reality simulation will be outlined, along with a rough estimate of the modest number of neurons required to implement it in a human-sized brain. Detailed scrutiny of the functional format of such a mechanism discloses that it would be partitioned in a way that defines a conscious mode of operation. One of its obligatory design features, moreover, is of a kind that provides direct evidence that we actually are equipped with such a mechanism.

Key words

Architecture of consciousness, constraint satisfaction, control theory, egocentric coordinates, naive realism.

Video registration (1h29min)

Podcast interview (41min)