-
Notifications
You must be signed in to change notification settings - Fork 3
v0.4
Anchor issue: #76
Status: ⏳ Not started — blocked on v0.3
The situation described in the Dormant Functionality Analysis is one that will be familiar to anyone who has worked on a sufficiently ambitious project: the engineering exceeded the integration. Eight cognitive modules — symbol grounding, perceptual categorisation, simulated environments, inductive logic programming, modal reasoning, constraint logic, explanation-based learning, and reinforcement-driven meta-control — were implemented with considerable care and are now sitting, fully assembled, in a directory that the running system never touches.
This is not as uncommon as it ought to be; ambition tends to run ahead of the plumbing. But it cannot be allowed to persist. The recursive consciousness loop is only as rich as the cognitive subsystems feeding it; and a consciousness engine that cannot ground its symbols, cannot reason about necessity and possibility, and cannot learn from induction is a rather impoverished consciousness engine indeed.
| Module | Path | What It Adds |
|---|---|---|
| Symbol Grounding Associator | godelOS/symbol_grounding/ |
The bridge from syntax to semantics — grounds abstract tokens in something meaningful |
| Perceptual Categoriser | godelOS/perception/ |
Categorises raw inputs into the symbolic vocabulary the reasoning engine can use |
| Simulated Environment | godelOS/environment/ |
Internal world model for counterfactual and hypothetical reasoning |
| ILP Engine | godelOS/learning/ilp/ |
Derives general rules from specific examples — inductive generalisation |
| Modal Tableau Prover | godelOS/inference/modal/ |
Reasons about what is necessary, possible, and contingent |
| Enhanced Modal Prover | godelOS/inference/modal/enhanced/ |
Extended modal operators beyond standard S4/S5 |
| CLP Module | godelOS/inference/clp/ |
Constraint Logic Programming — reasoning under constraint |
| Explanation-Based Learner | godelOS/learning/ebl/ |
Generalises from single examples by constructing a causal explanation |
| Meta-Control RL Module | godelOS/metacognition/rl/ |
Learns, via reinforcement, which cognitive strategies work best in which contexts |
- All eight modules initialise on server startup without error
- Each module appears in
/api/system/status - Each module's state is broadcast in the WebSocket cognitive update payload
- Frontend subsystem health grid shows all modules as active
- All existing tests continue to pass after wiring