How the loop goes negative

The starter loop says: reflect on emotions, design actions, filter through Path A/B, execute, observe what helped. Start moving.

But movement isn't automatically good. Actions generate feedback, and feedback can reinforce the wrong thing. The loop goes negative when the observation step — "what helped?" — is captured by an emotional signal instead of reading structural results. You're still acting. You're still observing. You're observing the wrong thing.

The relief loop

graph LR
  A[Action] --> B[Brief emotional relief]
  B --> C[Relief fades]
  C --> D[Anxiety returns]
  D --> A

Action produces brief emotional relief. Relief fades. Anxiety returns. More action. This is Path A running through the loop — the person is acting, but the observation step reads relief instead of results. The 80-hour work weeks from the distortion entry are this loop in practice. It looks productive from the outside. From the inside, "observe what helped" returns "I felt better for a while" rather than "this moved something structurally."

The trap: the loop is producing an outcome — emotional regulation. That's real. It's just not building anything. And because it reliably produces relief, the observation step confirms it as working.

The retreat loop

graph LR
  A[Action] --> B[Bad outcome or discomfort]
  B --> C["Safety need strengthens"]
  C --> D[Action space shrinks]
  D --> E[Less movement]
  E -->|Confirms the fear| C

You try something. It goes badly — or just feels uncomfortable. The safety need says "see, I told you." The constraining meme tightens its grip. One failed attempt becomes evidence for the boundary you suspected. "I tried, it didn't work, I knew I couldn't."

The action did generate data. But the safety need reads it selectively — confirming what it already believed while discarding what doesn't fit. A single failure becomes proof of a permanent limitation. The loop tightens: each attempt that doesn't go perfectly makes the next attempt less likely. After enough cycles, the person stops acting altogether — and from the frozen position, it looks like wisdom rather than fear.

The escalation loop

graph LR
  A[Action] --> B[Success]
  B --> C[Higher stakes]
  C --> D[More emotional pressure]
  D --> E[Strategy distorts]
  E --> F["Action serves the pressure,\nnot the goal"]
  F --> A

You act and succeed. Success raises the stakes. Higher stakes increase the emotional charge on the next decision. The next round of actions is less clear-eyed, more driven by the need to maintain the trajectory. Each win makes the next one feel more necessary — and the observation step gets noisier, because the emotional pressure makes it harder to see whether the strategy is still compounding or just feeding its own momentum.

This is the "enough" problem in motion. You're acting and succeeding, and the success itself is making you less able to evaluate what the strategy actually costs. The person getting promoted, earning more, achieving more — who is also sleeping less, seeing less, narrowing more — is in this loop. Every success feels like validation. The structural costs are invisible because the emotional signal is so loud.

The oscillation loop

graph LR
  A["Action toward success"] --> B["Safety need fires"]
  B --> C[Retreat]
  C --> D[Frustration about retreating]
  D --> A

Action toward success. Safety fires. Retreat. Frustration about retreating. Action again. Safety fires again. This is the safety-success deadlock playing out over time rather than as a single frozen moment.

The person isn't stuck — they're cycling. From the outside it might even look like they're making progress, because they keep starting things. But each retreat costs credibility — with themselves and others — making the next attempt start from a worse position. Energy is spent. No net movement. The observation step alternates between "I need to be bolder" and "I need to be more careful," providing contradictory guidance that keeps the oscillation going.

The comparison loop

graph LR
  A[Action] --> B[Visibility]
  B --> C[Comparison with others]
  C --> D[Mimetic rivalry]
  D --> E["Action shifts from\nbuilding to competing"]
  E --> A

You started building something real. The action made you visible. Visibility introduced internal mediators — people close enough to your world to trigger rivalry. Now you're optimizing against them instead of for your own trajectory. The action is still there but it's serving a different function.

The observation step shifts from "did this compound?" to "am I ahead?" — and that's a question with no stable answer, because rivals are also moving. The structural value of the work may be declining while the competitive intensity increases. You're working harder and building less.

The common thread

All five loops share the same failure mode: the observation step is captured. The person is still acting, still reflecting, still observing — running the starter loop as described. But "observe what helped" is reading the wrong signal:

The right column is the Path B version of each observation. The left column is Path A capturing the same step. The difference isn't whether you're running the loop — it's what your observation step is tuned to.

This is why the Path A/B filter exists in the loop. Without it, you can execute every step faithfully and still accelerate the wrong pattern. The filter doesn't guarantee you'll avoid these loops — a strong enough emotional need can pass it undetected. But it's the mechanism that catches the observation step being captured, before the loop has time to entrench.