Human-Centric Functional Modeling (HCFM) represents cognition as navigation through a conceptual space, where nodes correspond to concepts that impose boundary conditions on possible cognitive states, and edges encode reasoning operators as admissible transitions. This graph forms a network of constraints on cognitive evolution. Rather than attempting to specify the unknowable content of cognitive states, the framework focuses on the reliable functional behaviors that persist under these constraints. This approach constitutes a ``science of not knowing'': a rigorous study of epistemic boundaries formalized as a constraint algebra, where ignorance is modeled as the excluded volume of conceptual space, and reliable cognition emerges from the invariants preserved across that ignorance.