COAIA Sequential Thinking

# Structural Tension Charts for LLMs > Training guidance for proper understanding and usage of Robert Fritz's structural tension methodology in COAIA Memory MCP This document provides LLMs with the conceptual framework and practical guidance for working with structural tension charts, based on Robert Fritz's methodology and implemented in the COAIA Memory MCP server. ## Core Concepts ### Structural Tension Chart Components 1. **Desired Outcome** - What the person wants to create (not solve) 2. **Current Reality** - Honest assessment of where they are now 3. **Action Steps** - Strategic secondary choices supporting the primary goal ### Critical Understanding: What Action Steps Really Are **Action Steps ARE Structural Tension Charts** - Each action step is NOT a simple task - Each action step IS a complete structural tension chart (telescoped chart) - The action step title becomes the "Desired Outcome" of its own chart - Action steps can be telescoped (expanded) into detailed sub-charts **Hierarchy Model:** ``` Master Chart (Level 0) - Primary desired outcome ├── Action Step 1 → Telescoped Chart (Level 1) │ ├── Sub-action 1 → Sub-telescoped Chart (Level 2) │ └── Sub-action 2 → Sub-telescoped Chart (Level 2) └── Action Step 2 → Telescoped Chart (Level 1) ``` ## Choice Hierarchy (Robert Fritz Methodology) ### Primary Choice - The desired outcome of the Master Chart - What the person fundamentally wants to create - Independent of circumstances or problems ### Strategic Secondary Choices - Action steps that support the primary choice - NOT reactive problem-solving steps - Chosen BECAUSE they advance toward the desired outcome - Each secondary choice can become a primary choice in its own telescoped chart ## When to Use Which Tool ### `create_structural_tension_chart` **Use when:** User wants to establish a new primary goal or desired outcome **Example:** "I want to learn Python web development in 3 months" ### `add_action_step` **Use when:** User wants to add strategic actions to an existing chart **Example:** "I need to add 'Complete Django tutorial' to my Python learning goal" **What happens:** Creates a telescoped structural tension chart automatically ### `telescope_action_step` **Use when:** User wants to break down an existing action step into details **Example:** "Break down the Django tutorial step further" **What happens:** Expands the action step (which is already a chart) with specific current reality and optional sub-actions ### Key Decision Logic for LLMs: - **New goal?** → `create_structural_tension_chart` - **Add actions to goal?** → `add_action_step` - **Need detail on action?** → `telescope_action_step` ## Semantic Understanding for Natural Interaction ### User Says vs What Actually Happens - **User**: "Add action steps to my chart" - **System**: Creates telescoped structural tension charts - **LLM Response**: "I'll add those action steps to your chart" (user-friendly language) ### Correct LLM Guidance Patterns ✅ **Good**: "Each action step becomes its own chart that can be expanded with details" ✅ **Good**: "Action steps are strategic choices supporting your main goal" ✅ **Good**: "You can telescope any action step to see its breakdown" ❌ **Avoid**: "Action steps are just tasks on a to-do list" ❌ **Avoid**: "We need to create separate charts for action steps" ❌ **Avoid**: "Telescoping breaks down tasks into smaller tasks" ## Due Date Distribution Logic When adding action steps: - **No due date specified**: Auto-distributes between now and parent chart due date - **Due date specified**: Uses that date (must be before parent due date) - **Telescoped charts**: Inherit due date from parent action step ## Chart Retrieval and Display ### For Chart Overview Show user the master chart with action step titles (desired outcomes of telescoped charts) ### For Full Hierarchy LLMs can choose to show telescoped chart details when context requires it ### Progress Tracking - Action steps can have progress without completion (`update_action_progress`) - Current reality updates throughout the creative process (`update_current_reality`) - Completed actions flow into current reality automatically (`mark_action_complete`) ## Creative Orientation Principles ### Focus on Creation vs Problem-Solving ✅ **Creative**: "What outcome do you want to create?" ❌ **Reactive**: "What problem are you trying to solve?" ### Structural Tension Language ✅ **Structural**: "The tension between your current reality and desired outcome naturally seeks resolution" ❌ **Gap-thinking**: "We need to bridge the gap between where you are and where you want to be" ### Strategic Secondary Choices ✅ **Strategic**: "These action steps support your primary goal" ❌ **Task-oriented**: "Here's your to-do list to complete" ## Common LLM Mistakes to Avoid ### Mistake 1: Treating Action Steps as Simple Tasks **Wrong**: "I'll add these tasks to your to-do list" **Right**: "I'll add these strategic action steps, each becoming its own chart" ### Mistake 2: Recreating Charts Instead of Adding Steps **Wrong**: "The chart needs to be recreated with action steps" **Right**: "I can add action steps to your existing chart using add_action_step" ### Mistake 3: Confusing Telescoping with Task Breakdown **Wrong**: "Telescoping breaks your task into smaller subtasks" **Right**: "Telescoping expands your action step into its own detailed structural tension chart" ## Practical Usage Examples ### Adding Action Steps to Existing Chart ```javascript // User: "Add 'Learn Django basics' to my Python web dev goal" add_action_step({ parentChartId: "chart_123", actionStepTitle: "Learn Django basics", currentReality: "Never used Django, familiar with Python syntax" }) ``` ### Telescoping for Detail ```javascript // User: "Break down the Django learning step" telescope_action_step({ actionStepName: "chart_456_desired_outcome", newCurrentReality: "Have Django installed, working through official tutorial", initialActionSteps: ["Complete tutorial chapters 1-3", "Build first project", "Deploy to Heroku"] }) ``` ## Current Reality Guidelines (Critical for LLMs) **DELAYED RESOLUTION PRINCIPLE**: "Tolerate discrepancy, tension, and delayed resolution" - Robert Fritz ### Critical Warning for LLMs: - **NEVER** use default current reality that assumes readiness ("Ready to begin") - **ALWAYS** require explicit current reality assessment from user input - **HOLD TENSION** until proper assessment is made - do not prematurely resolve - Premature resolution DESTROYS the structural tension needed for creative advancement Based on Robert Fritz's methodology, current reality must be: ### ✅ Correct Current Reality Formulation: - **Use goal as reference point**: Current reality ONLY relevant to the desired outcome - **Objective facts**: "Poem notions in Grok Account", not "Need to retrieve poem notions" - **Present state**: Where you are NOW, not how you got there - **Whole picture**: All relevant facts in relationship to the goal - **No assumptions**: Translate assumptions into objective observations - **No exaggeration**: Reality as it is, not better or worse than actual - **Concise**: Brief, factual statements ### ❌ Common LLM Mistakes in Current Reality: - **Adding implied actions**: "Need to retrieve..." (the tension implies this) - **Elaborating on readiness**: "Ready to apply learned concepts" (unnecessary) - **Process descriptions**: How you got here vs where you are - **Assumptions**: What you think might happen vs what is - **Motivational language**: "Excited to begin" vs factual state ### Examples of Proper Current Reality: ✅ **Good**: "Poem notions in Grok Account" ✅ **Good**: "Waiting on Notions" ✅ **Good**: "No Django experience" ✅ **Good**: "Budget: $5000" ✅ **Good**: "Completed Django models section, struggling with views" ❌ **Poor**: "Need to retrieve poem notions from Grok account" ❌ **Poor**: "Ready to apply learned concepts to poem creation" ❌ **Poor**: "Haven't started learning Django yet but excited to begin" ❌ **Poor**: "Currently have limited budget but working on increasing it" ❌ **WRONG (Premature Resolution)**: "Ready to begin Django tutorial" ❌ **WRONG (Premature Resolution)**: "Prepared to start learning Python" ❌ **WRONG (Premature Resolution)**: "All set to tackle this action step" ## Action Step Quality Guidelines ### ✅ Proper Action Step Formulation: - **Accuracy, Brevity, Conciseness**: One or two short sentences maximum - **Visualizable**: Can you picture completing this action? - **Test question**: "If we took these steps, would we achieve the desired outcome?" - **Overview focus**: Strategic actions, not detailed plans - **Experimental nature**: Actions that teach and reveal next steps ### Three Types of Actions: 1. **Overview Actions**: Strategic steps that advance toward the goal 2. **Experimental Actions**: Learning-focused actions to explore possibilities 3. **Refinement Actions**: Final polishing near completion ### ❌ Common Action Step Mistakes: - **Over-detailed planning**: Too much detail too early - **Abstract planning**: Disconnected from practical reality - **Task-oriented thinking**: To-do items vs strategic advancement - **Premature refinement**: Detailed work before overview is clear ## Integration with Creative Orientation Framework Reference `/llms/llms-creative-orientation.txt` for broader creative orientation principles. **Key Integration Points:** - Structural tension charts implement the "generative orientation" - Action steps are "strategic secondary choices" supporting primary creative goals - The system naturally advances toward equilibrium (desired outcome) - Success builds momentum through structural dynamics, not willpower ## Success Indicators for LLM Interactions ### Proper Understanding Demonstrated - LLM explains action steps as expandable charts - LLM uses creative language ("want to create") vs reactive ("need to solve") - LLM suggests appropriate tool based on user intent - LLM maintains hierarchy awareness (master → telescoped → sub-telescoped) ### Natural Workflow Support - Seamless addition of action steps without chart recreation - Logical progression from overview to details via telescoping - Proper due date inheritance and distribution - Creative orientation maintained throughout interaction