deciqAI Knowing and Doing as One
Overview
When you say "I know I should but I just don't do it," the standard explanation is a failure of willpower. This skill rejects that. The gap between knowing and doing is the diagnostic signature that the knowing was never complete in the first place.
Wang Yangming (王阳明, 1472–1529) formalized this in 知行合一 (zhīxíng héyī): "There has never been one who knows but does not act. To know and not act is simply not to know." Modern research (Gollwitzer 1999, Pfeffer & Sutton 2000, Ries 2011) has independently converged on the same operational conclusion.
Composes with: metacognition · founder-mindset · lean-startup · critical-thinking
When to Use
Apply when: "I know I should but I haven't" · analysis paralysis going in circles · assessing whether someone actually learned from past experience · strategy/lessons-learned doc produced no operational change · org repeating the same mistake · someone invokes "say-do gap," "knowing-doing gap," or "Wang Yangming."
When NOT to use: barrier is demonstrably external (capital, regulation, waiting for a trigger) · "strategic patience" with specifiable duration · tempted to use it to shame rather than diagnose.
Coaching Novices (Adaptive Front Door)
- Engine mode: user has a concrete "I know but I haven't" case → run The Process directly.
- Coach mode: user is unfamiliar or has no concrete case → guide step by step.
In Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.
- One-line what-it-is: when you say "I know I should but I haven't," that's not a willpower problem — it's evidence that some specific part of knowing hasn't actually happened yet.
- Check fit against When to Use / When NOT to use. If the barrier is genuinely external, point elsewhere.
- Elicit their real case — a concrete "I know but I haven't," their own or someone they're helping. Don't run on hypotheticals.
[WAIT — do not advance until user responds]
- Walk through The Process one step at a time with their actual input.
[WAIT — do not advance until user responds]
- Close by naming the specific knowledge component that was incomplete, and the one next action that would test whether it is now complete.
[WAIT — do not advance until user responds]
The Process
Run the Knowing-Doing Audit. Find the gap, find what specifically isn't actually known, find the next acting step.
- State the claim of knowing — one sentence, not a slogan.
- State the concrete action not yet taken — what does doing-the-thing look like this week?
- State the gap — what has happened that is not the doing (thinking, talking, planning)?
- Apply Wang's diagnostic: what specifically would have to be additionally known for the doing to follow? Common missing: felt consequence; concrete picture of doing; alternative being protected; real motivations.
- Test at small scale — complete one missing component via small action. If still stuck, recurse.
- Recurse — each cycle reveals further missing components; terminates when knowing is sufficient.
- Audit others through doing, not words — evidence of learning = what they have done since.
- Stop-rule: audit ending with "I need to think more" has failed — must terminate in a concrete next action.
Output: the Knowing-Doing Audit
# Knowing-Doing Audit: <claim of knowing>
Claim: <one sentence>
Concrete action not taken: <what doing looks like this week>
Gap: <what happened that is not the doing>
Missing knowledge components: <component 1> / <component 2> / <component 3>
Test: <small action to complete one missing component>
Next action + date: <specific doing, specific time>
How I'll know it worked: <observable outcome or new component revealed>
→ Method in Action: Wang Yangming's Longchang Enlightenment, 1508
Pack: Knowing-Doing-as-One in Modern Practice
- Founder/startup: "Do things that don't scale" — knowing what customers need is constituted by personally delivering, not reading research.
- Hiring: Past behavior >> past explanations. Reference check: "what specific actions showed the learning?"
- Strategy: Plans with no change in calendar/budget/headcount within ~90 days are vocabulary, not strategy.
- Retros: Same issues quarter after quarter = lessons never learned because no action consistent with them was taken.
Applying It Well
- Diagnose ("what's missing?") not weaponize ("you don't know"). Apply to yourself first.
- Recursion terminates — if not after several cycles, the constraint is external.
- Fluency of retrospective narration = fluency, not learning.
→ Primary sources: references/sources.md
Common Rationalizations
[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.
| Fake move | Reality |
|---|---|
| [D] "I know what to do, I just lack motivation/willpower" | Name the specific knowledge component missing, complete it via small action, observe whether doing follows. The deficit is almost always knowing, not motivation. |
| [D] "I just need more information before I act" | The Cheng-Zhu position Wang criticized. Leads to neither knowing nor action. Act small on the components gated on action. |
| [D] "We held a retro and learned the lesson" | A retro that doesn't change behavior produced vocabulary, not learning. Test = action consistent with the lesson since. |
| [D] "They explained what they learned really well" | Fluency of narration is not evidence of learning. Evidence is what they have done since. Critical for hiring. |
| [D] Using this skill to shame rather than diagnose | Frame: "what specific component hasn't completed?" — not "you don't really know." |
| [D] Treating the audit as one-shot | Iterative. If doing still doesn't follow, run another cycle. The recursion terminates. |
| [D] Using the framework when constraint is genuinely external | Diagnose external constraints first or the audit produces false-positive "I don't really know" conclusions. |
| → Add [O] entries here after each real use — paste the actual failure pattern | What went wrong and why |
Red Flags
- "I know I should but I haven't" treated as willpower, not knowing gap
- Retrospective doc that changed no operational practice within 30–90 days
- Candidate's learning narrative far more sophisticated than post-learning behavior
- Team with 3+ retros on same root cause, no behavioral change
- Decision paralysis with remedy of "more analysis"
- "Strategic patience" invoked with no specifiable duration
Verification
- Claim of knowing stated as one concrete sentence
- Concrete action specified for this week
- Gap (what happened that is not the doing) honestly described
- Specific knowledge component named — not "willpower"
- Small-scale test action defined
- Next action with specific date on calendar
- Audit ended in action, not "I need to think more"
- When applied to others, evidence is behavioral not narrative
Part of deciqAI Knowledge Skills — 237 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. See it run → https://www.deciqai.com/s/knowing-and-doing-as-one · Built by deciqAI · github.com/deciqAI · Contributions welcome.
Agents: latest version & machine-readable metadata → https://www.deciqai.com/s/knowing-and-doing-as-one.json