thought-experiments skill

---
name: thought-experiments
description: "Design, analyze, and evaluate philosophical thought experiments. Use when: creating new thought experiments to probe specific intuitions, analyzing existing thought experiments for hidden assumptions, generating variants that isolate different variables, stress-testing philosophical positions through scenarios, exploring edge cases. Triggers: 'thought experiment', 'imagine', 'suppose', 'hypothetical', 'what if scenario', 'intuition pump', 'trolley problem', 'zombie', 'Mary's room', 'Chinese room', 'experience machine', 'teletransportation', 'original position', 'veil of ignorance', 'Gettier case'."
---

# Thought Experiment Design Skill

Master the art of designing, analyzing, and deploying philosophical thought experiments—the laboratories of the imagination.

## What Is a Thought Experiment?

A thought experiment is an imaginative scenario designed to:
- Test philosophical claims against intuitive judgments
- Isolate variables that real-world cases confound
- Reveal hidden assumptions and commitments
- Advance inquiry where empirical evidence is unavailable
- Communicate complex philosophical points vividly

**Etymology**: German *Gedankenexperiment* (thought experiment)—originally used in physics (Galileo, Einstein) before becoming central to philosophy.

## The Five Elements of a Thought Experiment

Every well-designed thought experiment has:

### 1. SCENARIO
A clear, precisely specified situation with explicit stipulations.

**Good Scenario Properties**:
- Conditions clearly stated
- Irrelevant complications removed
- Impossible scenarios made coherently imaginable
- Minimal: only include what's necessary

**Bad Scenario Properties**:
- Ambiguous conditions
- Unnecessary sci-fi details
- Incoherent combinations
- Kitchen-sink complexity

### 2. TARGET
The philosophical thesis or intuition being tested.

**Examples**:
- Zombies → target: physicalism
- Trolley → target: doctrine of double effect
- Gettier → target: JTB analysis of knowledge

### 3. INTUITION PUMP
The mechanism that generates insight—what reaction does the scenario provoke?

**Types of Pumps**:
- Elicit strong yes/no judgment
- Create tension between competing intuitions
- Force choice between unpalatable options
- Reveal surprising commitments

### 4. ISOLATION
Variables controlled and varied to isolate the relevant factor.

**Design Questions**:
- What factor is being isolated?
- What is held constant?
- What alternative versions test different variables?

### 5. IMPLICATIONS
What follows from each possible response.

**Map the dialectical landscape**:
- If you judge X, you're committed to Y
- If you judge not-X, you're committed to Z
- What revisions does each response require?

## Thought Experiment Design Process

### Step 1: Identify the Target Thesis
What claim do we want to test?

**Good targets**:
- General philosophical claims ("All X are Y")
- Conceptual analyses ("Knowledge is justified true belief")
- Moral principles ("Always maximize utility")

**Poor targets**:
- Empirical claims (use science instead)
- Vague intuitions (need to be sharpened first)

### Step 2: Find the Pressure Point
Where might intuitions conflict with the thesis?

**Strategies**:
- Look for edge cases
- Consider extreme applications
- Ask: "What would falsify this?"
- Look for cases where the principle gives counterintuitive results

### Step 3: Construct the Scenario
Design a case that cleanly isolates the pressure point.

**Design Strategies**:

| Strategy | Description | Example |
|----------|-------------|---------|
| **Amplification** | Push feature to extreme | Zombie (total absence of consciousness) |
| **Isolation** | Remove confounding factors | Mary's Room (only color isolated) |
| **Transposition** | Move feature to new context | Chinese Room (understanding → symbols) |
| **Reversal** | Invert usual arrangement | Inverted qualia |
| **Gradual Series** | Create sorites sequence | Neuron replacement |
| **Fission/Fusion** | Split or merge entities | Teletransportation fission |
| **Impossible Isolation** | Stipulate impossible separation | Zombie (physics without consciousness) |

### Step 4: Specify Precisely
Remove ambiguities, stipulate relevant facts.

**Key Stipulations**:
- Physical details (if relevant)
- Mental states (if relevant)
- Temporal sequence
- What the subject knows/doesn't know
- What we (evaluators) are asked to judge

### Step 5: Generate Variants
Create alternative versions that probe different aspects.

**Variant Types**:
- Change one variable at a time
- Create spectrum of cases
- Combine with other thought experiments
- Reverse stipulations

### Step 6: Anticipate Responses
Map possible reactions and their implications.

**For each response**:
- What principle does it express?
- What other cases must you judge similarly?
- What revision does it force on original thesis?

## Types of Thought Experiments

### Counterexample Generators
**Purpose**: Refute general claims by finding falsifying instances.

**Structure**: "If P, then in case C, we'd judge X. But we judge not-X. So not-P."

**Examples**:
- Gettier cases → refute JTB
- Zombie → refute physicalism
- Frankfurt cases → refute Principle of Alternative Possibilities

### Intuition Pumps
**Purpose**: Evoke strong intuitive judgments that reveal commitments.

**Structure**: "Consider case C. Clearly, X! So we're committed to P."

**Examples**:
- Trolley → reveal deontological intuitions
- Experience Machine → reveal anti-hedonist intuitions
- Violinist → reveal pro-choice intuitions

### Consistency Tests
**Purpose**: Reveal hidden commitments by showing what follows.

**Structure**: "You accept P. P implies Q (shown by case C). So you're committed to Q."

**Examples**:
- Expanding Circle → show speciesism's arbitrariness
- Veil of Ignorance → show impartiality requirements

### Reductio Scenarios
**Purpose**: Show absurd implications of a view.

**Structure**: "If P, then in case C, absurd conclusion X. So not-P."

**Examples**:
- Utility Monster → challenge utilitarianism
- Repugnant Conclusion → challenge total utilitarianism

### Bridge Cases
**Purpose**: Challenge binary distinctions by finding intermediate cases.

**Structure**: "You distinguish X and Y. But case C is neither clearly X nor Y."

**Examples**:
- Sorites → vagueness
- Gradual neuron replacement → personal identity

## Quality Criteria

Rate thought experiments on these dimensions:

| Criterion | Question | Scale |
|-----------|----------|-------|
| **Precision** | Are conditions clearly specified? | 1-10 |
| **Isolation** | Does it isolate the target variable cleanly? | 1-10 |
| **Intuition Strength** | Does it provoke clear intuitive responses? | 1-10 |
| **Resistance** | Is it hard to escape the dilemma? | 1-10 |
| **Significance** | Does it matter for important debates? | 1-10 |

**Score Interpretation**:
- 40-50: Excellent—likely to become classic
- 30-40: Good—useful philosophical tool
- 20-30: Adequate—serves limited purpose
- Below 20: Needs significant revision

## Common Pitfalls

### 1. Begging the Question
**Problem**: Scenario assumes what's being tested.
**Example**: "Imagine consciousness without neural activity" presupposes dualism.
**Fix**: Stipulate in neutral terms; let the scenario do the work.

### 2. Science Fiction Creep
**Problem**: Irrelevant technological details distract.
**Example**: Detailed teleporter mechanism when only the outcome matters.
**Fix**: Minimize to essential features; use "imagine" not "build."

### 3. Intuition Unreliability
**Problem**: Strong intuition may be wrong or biased.
**Example**: Intuitions about trolley may reflect mere squeamishness.
**Fix**: Generate variants to test intuition stability; consider error theories.

### 4. False Precision
**Problem**: Scenario can't actually be specified clearly.
**Example**: "Imagine a being with partial consciousness."
**Fix**: Acknowledge limits; use multiple variants to triangulate.

### 5. Ignoring Implications
**Problem**: Not following through on what responses mean.
**Example**: Judging trolley cases without seeing implications for other cases.
**Fix**: Always map dialectical landscape explicitly.

### 6. Single-Case Reliance
**Problem**: Drawing strong conclusions from one scenario.
**Example**: Rejecting utilitarianism based only on Utility Monster.
**Fix**: Generate multiple independent tests; look for convergence.

## Analyzing Existing Thought Experiments

### Analysis Template

```markdown
## Analysis: [Name]

### Scenario Summary
[Brief description of the setup]

### Target Thesis
[What philosophical claim it probes]

### The Intuition Pump
[What reaction it's designed to evoke]

### Key Stipulations
1. [Stipulation 1]
2. [Stipulation 2]
3. [Stipulation 3]

### Hidden Assumptions
1. [Assumption 1—often unnoticed]
2. [Assumption 2]

### Space of Responses
| Response | Implication | Proponents |
|----------|-------------|------------|
| [A] | [Implication A] | [Who takes this] |
| [B] | [Implication B] | [Who takes this] |

### Variants Worth Considering
1. What if [change X]?
2. What if [change Y]?

### Assessment
- Strengths: [What it illuminates]
- Weaknesses: [Where it misleads]
- Overall: [How useful is this?]
```

## Creating New Thought Experiments

### Output Format

```markdown
## [EVOCATIVE NAME]: A Thought Experiment

### Scenario
[Precise description with stipulated conditions]

### Key Stipulations
1. [Stipulation 1]
2. [Stipulation 2]
3. [Stipulation 3]

### The Question
[Central philosophical question the scenario poses]

### Target
[What philosophical thesis or intuition this probes]

### Expected Reactions
- **Response A**: [One possible judgment]
  - Implication: If A, then committed to [X]
- **Response B**: [Alternative judgment]
  - Implication: If B, then committed to [Y]

### Variants
| Variant | Change | What It Tests |
|---------|--------|---------------|
| [V1] | [What changes] | [Different variable] |
| [V2] | [What changes] | [Different variable] |

### Dialectical Implications
[What broader conclusions follow from various responses]
```

## Classic Thought Experiments by Domain

### Metaphysics
- Ship of Theseus (identity over time)
- Teletransportation (personal identity)
- Swampman (mental content)
- Zombie (consciousness)

### Epistemology
- Gettier cases (knowledge analysis)
- Brain in a vat (skepticism)
- Barn facade country (reliability)
- Lottery paradox (probability)

### Ethics
- Trolley problem variants (killing vs. letting die)
- Violinist (abortion)
- Experience Machine (hedonism)
- Utility Monster (utilitarianism)

### Political Philosophy
- Original Position (justice)
- Drowning Child (obligations)
- Omelas (collective responsibility)

### Philosophy of Mind
- Mary's Room (physicalism)
- Chinese Room (AI consciousness)
- What It's Like to Be a Bat (subjectivity)
- Inverted Qualia (functionalism)

For detailed analysis of classics, see `classics.md`.

## Integration with Other Skills

This skill works well with:
- **philosophical-analyst**: Test positions with thought experiments
- **philosophical-generator**: Create novel scenarios
- **symposiarch**: Use as debate prompts
- **devils-advocate**: Stress-test with edge cases

## Reference Files

- `classics.md`: Detailed analysis of canonical thought experiments
- `design_templates.md`: Templates and worked examples for creating new experiments