Self-Application as the Common Source of Benefit and Failure: A Control-Theoretic Model of Self-Regulation, Its Escapes, and a Standing-Cost Budget

We study a deterministic five-state control model of a self-regulating system — a generator of self-content, a demand/reduction process, and an attentional control loop — whose failure mode is collapse of the generator into a self-maintaining state with no outward pull, a structure taken from the phenomenology of depersonalization-derealization (DPDR). The artifact contains the paper's complete underlying set: the frozen model and its opt-in variants (permissive, regulator, window, generational), the committed experiment drivers with their pre-registrations and the regression test suite, the standalone substrate scripts, the cached numerical results behind every number, the figures, and the claim-by-claim register audit. In one concrete setting the paper locates the model's boundaries numerically, including a floor critical for escape, a monitoring-cost critical, and a hyperbolic capability-recoverability boundary; it further shows the same floor principle recurring in an SLD-resolution engine, in cut-elimination complexity, and in Datalog provenance. The deposit license is CC-BY-4.0 (the paper and figures); the code in this repository is additionally licensed MIT. AI assistance: the research and its content originate with the author — the question, the model, the design decisions, the interpretations, and every claim are the author's, and the model is the author's formalization of the author's own first-person account. Large language model tools were used instrumentally (through a custom agent harness written for the project), for implementing the simulator, running the experiments, drafting prose, and literature search — not for producing research content. Every number is recomputed from the frozen model by an executable checker (dpdr/reproduce.py, 128 checks, non-zero exit on mismatch). The author takes full responsibility for the contents.