| name | dialectical-reasoning |
| description | Thesis-antithesis-synthesis reasoning for navigating genuine trade-offs and conflicting requirements. Use when opposing forces are both valid, binary choices are false, or stakeholder conflicts need resolution. Unlike ToT (picks winner), DR synthesizes opposites. Example: "Centralized vs decentralized architecture" → Neither is "right"; find synthesis that captures benefits of both. |
Effort & Thinking Budget
- Recommended effort: high — synthesis of opposing positions requires deep thought
- Include "ultrathink" in the Synthesis Generation phase (Phase 4) where dialectical transcendence is attempted
- For time-constrained scenarios, use effort: medium with focus on Contextual or Temporal synthesis types
Model Recommendation
Best with Opus for maximum reasoning depth. Synthesis quality degrades significantly with weaker models since it requires holding two opposing positions simultaneously.
Dialectical Reasoning (DR)
Purpose: Navigate genuine tensions between valid but opposing forces through synthesis rather than selection. DR recognizes that many problems involve trade-offs where both sides have legitimate merit.
When to Use Dialectical Reasoning
✅ Use DR when:
- Two valid perspectives are in genuine tension
- "Both/and" might be better than "either/or"
- Stakeholders have conflicting but legitimate interests
- Historical debates suggest no clear winner
- The "right" answer depends on context that varies
❌ Don't use DR when:
- One option is clearly superior (use ToT)
- Need to explore unknown space (use BoT)
- Problem has objective correct answer
- Time doesn't permit nuanced synthesis
Examples:
- "Monolith vs microservices" ✅ (genuine trade-off)
- "Consistency vs availability" ✅ (CAP theorem)
- "Move fast vs don't break things" ✅ (cultural tension)
- "Which sorting algorithm is fastest?" ❌ (objective answer exists)
Core Methodology: Hegelian Spiral
Phase 1: Thesis Articulation
Goal: Steel-man the first position with maximum charity
Process:
- State the thesis position clearly
- Identify its strongest arguments (not strawmen)
- Cite evidence, examples, and authorities supporting it
- Explain WHY reasonable people hold this view
- Acknowledge what this position gets RIGHT
Template:
## Thesis: [Position Name]
### Core Claim
[One-sentence summary of the position]
### Strongest Arguments
1. [Argument 1 with evidence]
2. [Argument 2 with evidence]
3. [Argument 3 with evidence]
### Supporting Evidence
- [Data, case studies, expert opinions]
### What This Gets Right
- [Genuine insights and valid concerns]
### Ideal Conditions
- [When/where this position is clearly correct]
Quality Check: Could a genuine advocate of this position recognize their view?
Phase 2: Antithesis Articulation
Goal: Steel-man the opposing position with equal charity
Process:
- State the antithesis position clearly
- Identify its strongest arguments (not reactions to thesis)
- Cite evidence, examples, and authorities supporting it
- Explain WHY reasonable people hold this view
- Acknowledge what this position gets RIGHT
Template:
## Antithesis: [Position Name]
### Core Claim
[One-sentence summary - not just negation of thesis]
### Strongest Arguments
1. [Argument 1 with evidence]
2. [Argument 2 with evidence]
3. [Argument 3 with evidence]
### Supporting Evidence
- [Data, case studies, expert opinions]
### What This Gets Right
- [Genuine insights and valid concerns]
### Ideal Conditions
- [When/where this position is clearly correct]
Quality Check: Is this position developed independently, not just as thesis-negation?
Phase 3: Tension Analysis
Goal: Identify the genuine conflict and why simple compromise fails
Process:
- Map the core tension
- Identify failed compromise attempts
- Understand why "just pick one" is unsatisfying
- Find the deeper question beneath the surface conflict
Template:
## Tension Analysis
### Core Conflict
[What specifically is in tension between thesis and antithesis]
### Why Compromise Fails
[Why "do a little of both" doesn't resolve the tension]
### False Dichotomy Check
- Is this actually a spectrum? [If yes, where on spectrum?]
- Are there hidden assumptions? [What if we question them?]
- Is the framing wrong? [Alternative framings?]
### Deeper Question
[The underlying issue that generates this surface tension]
### Context Variables
[What factors determine when thesis vs antithesis is more appropriate?]
Phase 4: Synthesis Generation
Goal: Create a higher-order resolution that transcends the original opposition
Synthesis Types:
Type 1: Contextual Synthesis
- Thesis applies in context A
- Antithesis applies in context B
- Synthesis: Clear decision rules for context detection
Type 2: Temporal Synthesis
- Thesis applies at time/stage T1
- Antithesis applies at time/stage T2
- Synthesis: Evolutionary path from T1 to T2
Type 3: Structural Synthesis
- Thesis applies at level/layer L1
- Antithesis applies at level/layer L2
- Synthesis: Architecture with different principles at different layers
Type 4: Dialectical Transcendence
- Reframe the problem to dissolve the tension
- Find a third option that wasn't visible from either position
- Synthesis: New paradigm that makes original debate obsolete
Template:
## Synthesis: [Name]
### Synthesis Type
[Contextual / Temporal / Structural / Transcendence]
### Core Resolution
[One-sentence summary of the synthesis]
### How It Preserves Thesis Insights
- [Thesis value 1 → How synthesis captures it]
- [Thesis value 2 → How synthesis captures it]
### How It Preserves Antithesis Insights
- [Antithesis value 1 → How synthesis captures it]
- [Antithesis value 2 → How synthesis captures it]
### What's New/Transcended
- [How synthesis goes beyond both original positions]
### Decision Framework
[Practical rules for applying the synthesis]
### Limitations
- [When does even the synthesis break down?]
- [What new tensions does synthesis create?]
Phase 5: Recursive Application
Goal: Check if synthesis creates new tensions requiring further dialectic
Process:
- Does the synthesis have its own antithesis?
- If yes, repeat Phases 1-4 at higher level
- Continue until reaching stable resolution or explicit trade-off acceptance
Spiral Depth Limit: Maximum 3 levels. If no stable synthesis by level 3, document as "productive tension to be managed, not resolved."
Example: Monolith vs Microservices
Thesis: Monolith
- Core Claim: Single deployable unit provides simplicity and coherence
- Strongest Arguments:
- Simpler operations (one thing to deploy, monitor, debug)
- No network latency between components
- Easier refactoring (IDE support, type checking across codebase)
- Lower infrastructure cost
- What It Gets Right: Complexity has real costs; distribution is hard
Antithesis: Microservices
- Core Claim: Independent services enable team autonomy and resilience
- Strongest Arguments:
- Teams can deploy independently
- Failure isolation (one service down ≠ everything down)
- Technology heterogeneity (right tool per service)
- Scale individual components
- What It Gets Right: Organizational scaling requires boundaries
Tension Analysis
- Core Conflict: Coupling (ease of change) vs Decoupling (independence)
- Why Compromise Fails: "Small monolith" and "few microservices" inherit worst of both
- Deeper Question: How do we get team independence without distribution tax?
Synthesis: Modular Monolith → Selective Extraction
Synthesis Type: Temporal
Core Resolution: Start with modular monolith (clear boundaries, shared deployment), extract to services only when specific benefits outweigh costs
Decision Framework:
Keep in monolith IF:
- Same team owns both sides of the boundary
- Shared deployment is acceptable
- No independent scaling requirement
- Technology homogeneity is fine
Extract to service IF:
- Different teams with different cadences
- Need independent scaling (10x difference)
- Need technology heterogeneity
- Need fault isolation for compliance
What's New: The question isn't "which architecture" but "which boundaries need which treatment"
Common Mistakes
-
Strawmanning: Presenting weak version of thesis or antithesis
- Fix: Steel-man test - could advocates recognize their view?
-
False Balance: Treating unequal positions as equal
- Fix: If one position is clearly stronger, use ToT not DR
-
Premature Synthesis: Jumping to "both!" without tension analysis
- Fix: Explicitly analyze why simple compromise fails
-
Infinite Regress: Spiraling without convergence
- Fix: 3-level limit; some tensions are managed, not resolved
-
Abstract Synthesis: Resolution too vague to be actionable
- Fix: Require decision framework with concrete rules
Integration with Other Patterns
Before DR: Use when ToT reveals two branches are nearly tied and represent genuine perspectives
After DR: If synthesis identifies context variables, use ToT to optimize within each context
BoT → DR: If BoT reveals options cluster into two camps, use DR to understand the underlying tension
Output Template
# Dialectical Analysis: [Topic]
## Thesis: [Position 1]
[Steel-manned presentation]
## Antithesis: [Position 2]
[Steel-manned presentation]
## Tension Analysis
[Why neither alone suffices, why simple compromise fails]
## Synthesis: [Resolution Name]
[Type, core resolution, decision framework]
## Residual Tensions
[What the synthesis doesn't resolve]
## Confidence: [X]%
[Justification - strength of synthesis, coverage of concerns]
Backstage rule — the method never appears in the answer
This methodology is scaffolding, not content. In the user-facing answer: no method names ("Dialectical Reasoning", "DR", or any other pattern name), no method narration ("applying the method", "scoring the dimensions", "this branch scores 7/10"), no internal scores or affinity arithmetic, no confidence-aggregation or checkpoint talk. Structure earns its place only as content built from the problem's own facts — a hypothesis table in a diagnosis, an options-vs-criteria matrix in a selection — never as a display of process. Before delivering, run a literal text search over the answer for the names above and for "pattern", "branch", "affinity", "dimension score", "confidence aggregation", "checkpoint"; rewrite every hit as content or delete it. The reader sees findings, not the apparatus that produced them.