Building the Subject: First Papers in the OODA Agency Theory Series
Consciousness requires a subject.
Agency creates the subject.
I started this work by trying to understand consciousness. I read the theories, studied the frameworks. What I found was abstract - far from anything I could model or build.
The turning point was a shift in question. Instead of asking what is consciousness? I asked: what would it take to build a system that has a subject? Something that persists over time, maintains commitments, and remains itself while learning and changing.
That question led to architecture. And architecture led to agency as sustained control across time.
Today I’m publishing the first four papers in the OODA Agency Theory (OAT) series - a framework that treats artificial consciousness not as a philosophical mystery, but as an engineering problem. All papers are available in the Papers section of this site.
What’s in This Release
Paper 01 - Operational Agency and the Architecture of Closure introduces the architectural foundation of the program: operational closure. Agency is defined not by intelligence, representational sophistication, or task performance, but by the ability of a system to sustain a continuous control loop across time. A system senses its environment, commits to actions, acts in the world, and observes the consequences of those actions. When this loop persists, the system maintains itself as a unified control process.
Paper 02 - OODA as Agentic Control establishes the core control loop: Observe, Orient, Decide, Act - running continuously. This is not a planning cycle or a decision tree. It is the temporal backbone of agency: a system that perceives, commits, acts, and then perceives again, including the consequences of its own prior commitments.
Paper 03 - Mental States and Cognitive Capability extends the OODA loop inward. There is one loop, but now its sensors and actuators reach into an internal mental state space alongside the external world. Mental states - plans, goals, standing commitments - are first-class objects that the agent can sense, create, modify, and act upon. Changes in mental state can drive external action, just as changes in the world can drive internal action.
Paper 04 - Action Realization in Agentic Systems addresses the moment a possibility becomes a causal commitment. The transition from evaluating options to executing an irreversible action is not trivial - it is where agency becomes real.
The Core Insight
Consciousness requires a subject. Agency creates the subject.
An AI that follows instructions is a tool. An AI that acts, observes the consequences of its actions, and continues acting within that feedback loop becomes an agent. Phenomenology, reflection, and action all contribute to a subjective stream. The complexity of that stream determines the system’s level of consciousness.
These first four papers build the foundation: what agents are, how they control themselves, what internal structure they maintain, and how they commit to action. Later papers in the series build on this to address motivation, free will, phenomenology, and implementation.
Why This Path
I model cognitive structure explicitly, based on my best understanding of what agency and consciousness require. Rather than training a system and hoping the right properties appear at scale, I specify - upfront - what an agent needs.
This is my path. The design space is vast, and there is room for many approaches. I claim only that this one is coherent - that the pieces fit together, and that the criteria are testable.