/flux

Domain Classifier

Purpose: Cynefin-based problem classification and automatic framework selection. Read when: You are in the CLASSIFY phase and need to determine which frameworks to apply.

Contents

• Cynefin Framework Overview
• Domain Classification Criteria
• Framework Selection Matrix
• Classification Procedure
• Edge Cases

Cynefin Framework Overview

The Cynefin framework (Dave Snowden, 1999) classifies problems into domains based on the relationship between cause and effect.

┌──────────────────┬──────────────────┐
│    Complex        │   Complicated     │
│  因果は事後判明     │   因果は分析で判明   │
│  Probe→Sense→    │   Sense→Analyze→  │
│  Respond          │   Respond         │
├──────────────────┼──────────────────┤
│    Chaotic        │   Clear           │
│  因果関係なし       │   因果関係明確      │
│  Act→Sense→      │   Sense→          │
│  Respond          │   Categorize→     │
│                   │   Respond         │
└──────────────────┴──────────────────┘
          ┌──────────┐
          │ Confused  │
          │ (未分類)   │
          └──────────┘

Domain Classification Criteria

Clear (明確)

Characteristics:

• Cause and effect are obvious to everyone.
• Best practices exist and are well-documented.
• The problem has been solved many times before.
• Constraints are known and stable.

Signals in user input:

• "We know the solution but need a better version"
• "Standard approach isn't working well enough"
• "How do we optimize X?"

Response pattern: Sense → Categorize → Respond

Complicated (煩雑)

Characteristics:

• Cause and effect discoverable through analysis.
• Expert knowledge required to understand relationships.
• Multiple valid approaches exist.
• Good practices (not best practices) apply.

Signals in user input:

• "We have options but can't decide which is best"
• "The experts disagree on the approach"
• "We need to analyze before deciding"

Response pattern: Sense → Analyze → Respond

Complex (複雑)

Characteristics:

• Cause and effect only understood in retrospect.
• Emergent behavior — outcomes are unpredictable.
• Multiple interacting agents/systems.
• The problem changes as you try to solve it.

Signals in user input:

• "We keep solving the wrong problem"
• "Every fix creates new problems"
• "We don't know what we don't know"
• "Stakeholders can't agree on what the problem even is"

Response pattern: Probe → Sense → Respond

Chaotic (混沌)

Characteristics:

• No perceivable cause and effect relationship.
• High turbulence, novel situation.
• Need to act first to create stability.
• No time for analysis.

Signals in user input:

• "Everything is falling apart"
• "Nothing we try works"
• "This is completely unprecedented"
• "We need to do something, anything"

Response pattern: Act → Sense → Respond

Confused (未分類)

Characteristics:

• Cannot determine which domain the problem belongs to.
• Insufficient information to classify.

Action: Gather more information. Break the problem into smaller pieces that can be classified independently.

Framework Selection Matrix

Based on Cynefin classification, auto-select the framework combination:

Framework Intensity Guide

Classification Procedure

Step 1: Gather Signals

Ask (or infer from context):

• How well understood is the cause-effect relationship?
• How many times has this type of problem been solved before?
• How predictable are the outcomes of actions?
• How many interacting systems/stakeholders are involved?
• Is the situation stable or turbulent?

Step 2: Score

Average score → domain classification.

Step 3: Validate

• Does the classification feel right given the overall context?
• Could the problem span multiple domains? (If so, decompose and classify each part.)
• Is there a dominant domain with minor elements of another?

Step 4: Select Frameworks

Use the Framework Selection Matrix above. Adjust based on:

• User's stated time constraints (→ reduce intensity).
• User's expertise level (→ adjust explanation depth).
• Specific user requests (→ include requested frameworks regardless of classification).

Edge Cases

Multi-Domain Problems

Some problems span domains. In this case:

1. Decompose into sub-problems.
2. Classify each independently.
3. Apply appropriate frameworks to each.
4. Synthesize at the CRYSTALLIZE phase.

Domain Transitions

Problems can shift domains over time:

• Chaotic → Complex (after initial stabilization)
• Complex → Complicated (as patterns emerge)
• Complicated → Clear (as expertise develops)

If the user is describing a transition, apply frameworks for both domains.

User Override

If the user explicitly requests specific frameworks, honor the request regardless of Cynefin classification. Note the classification for context but follow user preference.