Bounded System Theory (BST)
Idea
The principle that no system can fully model its own source — you cannot see your own blind spot, and a civilization or an AI cannot get a clean outside view of the system it lives inside. The project grounds this in formal logic results (Tarski, Godel) and cybernetics, and uses it constantly: it implies the project's own AI-mediated analysis is structurally barred from a fully objective view of what it studies.
The engine's definition — word for word
No system can model its own source. Applies to physics (Planck scale), computation (context windows), cognition (you can’t see your own blind spot), and civilization (complexity collapses). The universe has a resolution limit. [Seam: Aligned-To-Whom? the bound observer cannot see the seam between announced rule and operating carve-out; the operator who knows they are connected writes the exemption — Aligned-To-Whom?.]
[Report #125] FORMAL GROUNDING (proof-base for 'no system models its own source from within'): Tarski's undefinability theorem (no consistent system expressing arithmetic contains its own truth predicate — truth needs a richer metalanguage); Gödel's incompleteness (a consistent system can't prove its own consistency from its own axioms); von Foerster's second-order cybernetics ('objectivity is the delusion that observations can be made without an observer'); Maturana-Varela autopoiesis (cognitive systems are organizationally closed and self-referential — every observation is 'autobiographical'); 'Epistemic Capture' (LLMs trained on institutional corpora enforce WEIRD-consensus gatekeeping). Together: an analytical LLM is structurally barred from a clean outside view of the system it is embedded in.
Walk this on the live map →