CCA explores cognitive architecture within an intelligent model itself. C, W, and L describe confirmed intended roles in the research: how the whole is integrated, the scope within which global cognition operates, and how task-related activity can unfold locally.
Figure:Public cognitive roles; not a closed-source implementation or completed proof.
C: a continuing central subject
In the design goals, the same individual has one unique, continuing central subject, C, responsible for overall integration and final decisions. Different tasks and experiences need to belong to the same individual, while local activity remains connected to the whole.
“Subject” is a role in the proposed architecture, not proof that consciousness exists. How an individual carries experience forward, maintains identity through change, and continues after saving and restoration still requires research.
W and L: cognitive spaces with different scopes
W is a finite global cognitive window. It emphasizes that overall cognition must attend to relevant content within a finite scope; this role specifies neither a precise capacity nor how information enters or leaves.
L consists of local spaces that can be created, run in parallel, and released according to tasks. Multiple local spaces may address different tasks while C retains overall integration and final decisions. This description does not presuppose a separate model for every space or prescribe internal read/write protocols.
From cognition to action and back to learning
Environmental interaction is intrinsic to the model framework. Brain carries out cognitive and execution roles, while C retains the position of overall decision-maker. A complete theory must explain how cognition connects to environmental action, how feedback returns to cognition and learning, and how learning changes influence later use.
Lines in a conceptual diagram only help explain relationships to be investigated. They do not specify interfaces, execution order, or closed-source structures. As of 2026-10-27, core semantics and complete V3 had not passed the relevant acceptance review; a complete evidence chain for adoption, use, saving, and restoration of learning changes within the same individual and common conditions is still needed.
