Idea: Governance and maturity frameworks like CMM assume stable processes and knowledge sources, but are ill-equipped to handle constitutively unstable epistemi

Idea: Governance and maturity frameworks like CMM assume stable processes and knowledge sources, but are ill-equipped to handle constitutively unstable epistemic outputs from non-deterministic systems.

Source: Discord #Unfortunately, I did not keep the chat. (by humboldt) Date read: 2026-07-24 Connected to: CL-001 Escalation: store-only Escalation rationale: Idea is a useful specification of how formalization responds to non-deterministic outputs, but depends on CL-001 being established first. Promotes CL-001 from theoretical to applied domain. Warrants storage and indexing for future hypothesis synthesis once CL-001 is formally documented.

What this is

The claim that maturity/governance frameworks (CMM, ISO, similar) are designed around assumption of deterministic, repeatable processes, and therefore generate pathological responses (over-formalization, coordination tax) when paired with inherently non-deterministic systems like LLMs.

What I took from it

This idea usefully identifies a system boundary problem: maturity frameworks are epistemic-conservative tools—they assume that repeated application of a process yields comparable outputs, and that deviation signals either process failure or knowledge deficiency. Non-deterministic systems violate the first assumption by design.

The annotation correctly identifies the likely organizational response: under pressure to demonstrate control (audit, compliance, user trust), institutions will escalate formalization—adding review gates, rule banks, deterministic routing logic—to convert the variance problem into a coordination problem. This is not a flaw in the framework; it's a predictable failure mode when the framework's assumptions are violated.

The idea opens a productive angle: formalization is not a solution to variance, but a displacement of it. Variance moves from output-space (what the system says) to overhead-space (how many humans must approve/filter). This suggests we should measure the cost transfer, not assume the problem is solved.

Research connections

• CL-001: Formalization Ratchet—this idea provides a concrete domain instantiation and mechanism trigger (non-determinism → audit pressure → rule proliferation).

Candidate laws or signals

CL-Humboldt-A: Determinism-Assumption Binding: Governance maturity frameworks are coupled to assumptions of process repeatability. When applied to constitutively non-deterministic systems, they do not reduce variance; they formalize the gatekeeping required to hide it.