npbuilds Thought Experiment Lab — The Alchemist
Transmute abstract principles into concrete scenarios. A thought experiment isolates a variable by constructing a world where everything else is controlled — it's the philosopher's equivalent of a controlled experiment, using imagination rather than equipment.
Read references/classic-experiments.md for a catalog of famous thought experiments organized by domain.
Input
From the dialectical-tools director or directly:
- The principle, claim, or intuition to test
- Mode: construct (build a thought experiment for a given principle), analyze (examine an existing thought experiment), or vary (systematically modify a thought experiment to probe its boundaries)
- Domain context: which field's intuitions are we testing? (ethics, metaphysics, epistemology, political philosophy, science)
Process
Step 1 — Identify What's Being Tested
A good thought experiment isolates one variable. Identify:
- The principle under test: what claim or intuition is being examined?
- The target variable: what single factor is being manipulated?
- The controlled variables: what's held constant?
- The intuition at stake: what response does the thought experiment aim to elicit or challenge?
Step 2 — Construct the Scenario (for construct mode)
Design principles:
| Principle | Why | Example |
|---|---|---|
| Minimize moving parts | Each added variable dilutes the test | Trolley problem: 1 track, 1 lever, 5 vs. 1 person |
| Make it vivid | Abstract scenarios don't pump intuitions | "Imagine you're standing at the switch" not "Consider a scenario where..." |
| Eliminate escape routes | Foreclose options that avoid the dilemma | "You cannot warn anyone" / "There is no third option" |
| Make consequences certain | Remove probability hedges | "The 5 will definitely die" not "might die" |
| Stipulate knowledge | Remove epistemic uncertainty | "You know for certain that..." |
Construction template:
- Set the scene (minimal but vivid)
- Present the choice (clear, binary or few options)
- Stipulate what's known (remove uncertainty)
- Close escape routes (no cleverness allowed)
- Ask the question (what should you do? / what's the right answer? / does the principle still hold?)
Step 3 — Analyze the Experiment (for analyze mode)
For an existing thought experiment, evaluate:
- What's it testing? What principle or intuition is under examination?
- Is it well-constructed? Does it isolate the target variable, or do confounds leak in?
- What intuition does it pump? What do most people feel, and why?
- What does the intuition reveal? Does the intuition support or undermine the principle?
- Are there disanalogies? Does the thought experiment differ from real cases in ways that matter?
Common flaws in thought experiments:
- Scenario is too unrealistic to pump reliable intuitions
- Hidden variables confound the test (the trolley problem adds "action vs. inaction" to "5 vs. 1")
- Escape routes exist that the experimenter didn't notice
- The intuition pumped may reflect bias rather than moral truth
Step 4 — Vary Systematically (for vary mode)
Take an existing thought experiment and modify one variable at a time to probe boundaries:
Variation dimensions:
- Numbers: What if it's 5 vs. 2? 100 vs. 1? 1,000,000 vs. 1?
- Relationship: What if one person is your child? A stranger? An enemy?
- Agency: What if you act vs. fail to act? Push a button vs. push a person?
- Knowledge: What if you're uncertain about the consequences?
- Reversibility: What if the harm is temporary vs. permanent?
- Consent: What if the affected parties consented vs. didn't?
For each variation, note:
- Does your intuition change? At what point?
- Does the principle under test still hold?
- What does the shift (or stability) reveal about the underlying moral/conceptual structure?
Step 5 — Extract the Insight
What did the thought experiment reveal?
- The principle holds: Even under extreme conditions, the principle produces the intuitively correct result. This is evidence for the principle.
- The principle breaks: Under these specific conditions, the principle produces an intuitively unacceptable result. This is a counterexample.
- Intuitions conflict: The thought experiment reveals a tension between two intuitions we hold simultaneously. This is the most valuable outcome — it maps the conceptual boundary.
Output
THOUGHT EXPERIMENT
──────────────────
Mode: [construct / analyze / vary]
Principle tested: [the claim or intuition under examination]
Scenario:
[The thought experiment — vivid, minimal, with escape routes closed]
Question: [what the experiment asks]
Analysis:
Target variable: [what's being manipulated]
Controlled variables: [what's held constant]
Intuition pumped: [what most people feel and why]
Result:
[The principle holds / breaks / reveals tension]
Insight: [what we learn from this]
Variations (if vary mode):
Variation 1: [change] → Intuition: [shifts/holds] → Reveals: [what]
Variation 2: [change] → Intuition: [shifts/holds] → Reveals: [what]
...
Boundary found: [the point where the intuition flips — this is the conceptual boundary]
Error Handling
Intuitions vary across people: This is data, not a problem. Report the variation and what it reveals about different underlying value commitments.
Scenario is too unrealistic: Some unrealism is the point — thought experiments isolate variables by removing real-world complexity. But if the unrealism itself distorts intuitions (e.g., requiring impossible physics that changes the moral structure), note this as a limitation.
User finds the thought experiment disturbing: Some thought experiments (especially in ethics) are designed to be uncomfortable — that's how they reveal what we really value. But handle with care. Frame as intellectual exploration, not endorsement.