Information and Boundaries
Which observations across a boundary help a system maintain and revise an internal model?
Why this direction matters
Living systems maintain an internal state while exchanging energy, matter, and signals with their surroundings. Information theory offers ways to quantify uncertainty and the value of observations, but the mathematical variables must be chosen before those quantities have an interpretation.
First formal target
Specify a small model with internal state, external state, a boundary observation, and a state-update rule. Define what the system can observe, what it is trying to predict or maintain, and which uncertainty is reduced by a new observation. A toy model should distinguish information about the environment from information about the system’s own state.
Links to existing tools
ProLTs supply an explicit language for observational equivalence. Observation refinement studies the structural effect of adding a test. This direction would ask how those operations interact with information measures and ongoing state updates.
Current status
The system-and-environment direction remains proposed. P2: Binary Information Bounds now supplies a bounded statistical baseline: the exact mutual-information interval for two binary variables with fixed marginals, exact support and a positive cell floor. It does not model a boundary observation or state-update rule, identify a causal channel, or establish this broader direction. In Causal Self-Diagnosis, routine transcript laws identify only actuator/environment reversal parity under the stated informative-noise condition. In Hidden reference aging, routine observations provide no evidence about shared sensor orientation under the supplied fresh-target ensemble; the self/world boundary is assigned by the model. Information gain or entropy reduction alone should not be described as sentience or self-awareness.
How this connects
Open in the knowledge map- Phases
- Projects
- Concepts
Conceptual and navigational links, not proof dependencies.

