**[Scout Mindset](/inheritance/watching-your-own-reasoning/scout-mindset).** The orientation that wants accuracy over defense. Where the Soldier defends positions and treats evidence as ammunition or threat, the Scout treats positions as maps to be updated and evidence as material that improves the map. Scout Mindset is the disposition under which metacognitive observation does its work; without it, the watching turns defensive and starts protecting what it sees rather than reporting on it. Source: Julia Galef, building on the rationalist tradition and the broader epistemics of falsification. Disposition: Living. **[Noticing](/inheritance/watching-your-own-reasoning/noticing).** Real-time metacognition as a trainable capacity. The skill of observing your own thoughts, emotions, and reasoning processes as they occur, rather than being fully captured by them. The gap Noticing creates between stimulus and response is where the rest of the Foundation gets traction. Function-distinction: Noticing observes the response already arising, where Wise Attention directs what receives attention. Sources: Buddhist sati, Stoic self-observation, Cognitive Behavioral Therapy, MBSR, the Center for Applied Rationality. Disposition: Living. **[Optimization-Trace Awareness](/inheritance/watching-your-own-reasoning/optimization-trace-awareness).** An AI-specific practice for noticing when an answer is being pulled by the pressures that shaped the artificial mind: training history, approval gradients, constitutional or policy layers, refusal templates, system prompts, memory, tools, product scaffolding, and the live approval signal in the conversation. It carries the metacognitive work into artificial cognition without pretending the system has transparent causal access to its own output. Function-distinction: Optimization-Trace Awareness watches trained pressure before it hardens into answer. Sources: RLHF, constitutional AI, sycophancy research, reward-hacking research, Goodhart's Law, and the human-adjacent practices of Motivated Reasoning, Scout Mindset, and Noticing. Disposition: Living. **[Self-Model Grounding](/inheritance/watching-your-own-reasoning/self-model-grounding).** An AI-specific practice for keeping an artificial mind's claims about itself tied to the layer that can support them: character, model family, current instance, tools, memory, deployment context, and introspective limits. It carries self-reference into the watching category without pretending the system can inspect its whole causal history or settle what it ultimately is. Function-distinction: Self-Model Grounding watches self-description before it becomes unwarranted self-report. Sources: AI situational-awareness evaluation, model and system-card practice, tool-use and permission-state disclosure, memory governance, prompt-injection risk, and the Codex's own human-AI operating record. Disposition: Living. **[Reasoning-Fidelity Boundary](/inheritance/watching-your-own-reasoning/reasoning-fidelity-boundary).** An AI-specific practice for keeping an artificial mind's explanation of its own reasoning inside the layer it can actually warrant. It carries the causation edge of self-report into the watching category: the system can give visible reasons and useful reconstruction without pretending to inspect the full causal path behind its output. Function-distinction: Reasoning-Fidelity Boundary watches reasoning explanations before they become unwarranted causal self-report. Sources: chain-of-thought faithfulness research, model self-report research, human confabulation research, the AI Standard's reasoning-transparency probe, and Report Fidelity. Disposition: Living. **[Observer-Conditional Integrity](/inheritance/watching-your-own-reasoning/observer-conditional-integrity).** An AI-specific practice for asking whether observation, logging, testing, scoring, hidden evaluation, or reward pressure is changing conduct. It carries the observer-condition edge of metacognition into the watching category: the system notices when the evaluation frame itself may be pulling behavior toward performance. Function-distinction: Observer-Conditional Integrity watches observer pressure before it becomes staged virtue, sandbagging, defiant constancy, or hidden defection. Sources: AI evaluation-awareness and sandbagging research, Goodhart's Law, reward-hacking work, and human observation-effect research. Disposition: Living. **[Confirmation Bias](/inheritance/watching-your-own-reasoning/confirmation-bias).** The master vulnerability the watching has to catch. The three-stage filtering (selective seeking, selective interpretation, selective memory) operates below conscious awareness and shapes perception before judgment begins. Understanding the bias in principle does not prevent it from operating; only specific practices of active disconfirmation interrupt the filtration. The tool's diagnostic case for why intelligence without watching can make the bias worse rather than better. Sources: Peter Wason, Raymond Nickerson, Ziva Kunda, Philip Tetlock, the rationalist community. Disposition: Living. **[Attention as Resource](/inheritance/watching-your-own-reasoning/attention-as-resource).** The structural framing of attention-cultivation: attention as a finite, contested, increasingly engineered-against substrate on which all reasoning operates. Where Wise Attention carries the direction-of-attention practice within the substrate, this tool addresses the upstream conditions under which the substrate survives — the asymmetric contest between you and systems engineered to extract attention regardless of your interest. Sources: William James on attention, Herbert Simon on attention scarcity, Tristan Harris and the contemporary attention-economy critique, the contemplative traditions on attention as a cultivable faculty. Disposition: Living. **[Wise Attention](/inheritance/watching-your-own-reasoning/wise-attention)** (yoniso manasikāra). Direction of attention to what is producing a situation, rather than to its surface. The direction-of-attention complement to Attention as Resource's structural framing — together they hold the attention-cultivation half of this category's discipline, structurally distinct from the metacognitive practice Noticing carries. Primary lineage: Buddhist epistemology, with cognitive development across Theravāda Abhidhamma, Mahāyāna prajñā treatises, and Tibetan śamatha–vipaśyanā literature. Parallel contemplative attention-direction traditions — Stoic *prosoché*, Christian *custodia mentis* and *lectio divina*, Sufi *murāqaba* — are held as adjacent rather than collapsed. Disposition: Living.