cat ./research/metacognitive-analysis.md

Metacognitive Analysis

Clinical Psychology of the 16 Pivots

Three-stage clinical psychology analysis of the reflective functioning that enabled each pushback moment. Initial assessment: RF 6-7. Recalibrated to RF 6-9 after disclosure of 20 years meditation practice including 15 Zen sesshin. Three peaks at RF 8-9.

LLM PROVENANCE: Produced by Claude (Anthropic). Not independently verified. Starting material, nothing more.

Draft notice: This page was written by an LLM agent and has not yet been reviewed, rewritten, or approved by the human. It exists as raw material. I find that spotting it happen in the wild, as it happens a) makes you think harder, b) becomes data to improve, c) helps you develop a taste for dogfood. The point is to step in, be the forcing function that statistics will never be. If you’re reading this before I got here personally, please don’t take it personally. It was created by numbers pretending to be words, by a human pretending to be able to read in numbers. That said, all slop must die. To battle.

Context

After mapping the 16 pushback moments in The Fight Card, the question became: what kind of metacognitive ability enables a human to consistently identify when an LLM is being sycophantic?

This analysis uses Fonagy’s Reflective Functioning Scale (RF) as the primary framework, drawing on clinical and cognitive psychology (metacognitive monitoring, mentalization, decentering, epistemic vigilance, and a few others).

Three stages

Stage 1: Initial assessment

Based on the 16 pivot transcripts alone, the analysis placed the human operator at RF 6-7, “marked capacity for reflective functioning” with moments approaching RF 8. Six primary metacognitive constructs were identified. One gap was flagged: prospective metacognitive regulation.

Stage 2: Recalibration

The human disclosed 20 years of meditation practice including 15 Zen sesshin (intensive retreats, typically 7 days of 12+ hours daily sitting). This materially changed the assessment. The ceiling moved from RF 7 to RF 9, with three peaks:

  • Rounds 5-6-7 (RF 8): Sustained metacognitive monitoring under technical pressure
  • Rounds 14-15 (RF 8-9): Koan-like resolution, holding contradiction without forcing resolution
  • Round 16 (RF 9): Don’t-know mind under production pressure: the ability to say “I don’t know what this is yet” when the model was offering confident answers

Stage 3: Correction

The human pointed out that the “prospective regulation gap” was incorrect: since a specific commit (9960cd2), the engineering work itself was the prospective regulation. The governance systems mapped onto distributed cognition research in ways that weren’t planned, including Hutchins’ foundational work, which was, of all things, a study of naval navigation.

Full analysis (three reports) available in the research archive. This page is a summary.