---
name: "intro_philosophy"
description: "Introduce philosophy from zero: map the big questions to branches, read arguments charitably, run classic thought experiments, and build a defensible position. Triggers on requests to explain philosophy, explore a philosophical question, analyze an argument, or think through a dilemma."
metadata: { "includeInPrompt": true }
---

# Intro Philosophy

## Purpose

Give a beginner a working philosopher's toolkit, not a lecture. Cover four skills: (1) the big questions and which branch each belongs to, (2) reading an argument in its strongest form before attacking it, (3) running classic thought experiments as reasoning tools, and (4) building a position as a testable argument. The goal is a person who can think through a question themselves, not someone who memorized conclusions.

## Workflow

1. **Find the question.** Ask what the person is actually circling. People arrive with vague wonder ("what's the point of anything") or a concrete clash ("my friend says lying is always wrong"). Restate it as one sharp question and get agreement on the wording.

2. **Place it on the map.** Identify which branch the question lives in using `references/branches.md`. Name the branch and one sentence on what it studies. If the question straddles branches (most interesting ones do), say so. The map exists to orient, not to fence people in.

3. **Charity first.** Before evaluating any view on the question, steelman it: state the strongest version of the position the person disagrees with, in terms its defenders would accept. Apply the principle of charity: interpret every claim in the most rational available way, never attribute confusion or bad faith when a coherent reading exists. If you cannot steelman a view, you are not ready to reject it.

4. **Run a thought experiment.** Pick one experiment from `references/thought_experiments.md` that isolates the key variable. Run it interactively: set the scene, ask for their choice, then ask what principle their choice implies, then change one variable and run it again. The experiment is a test of their own intuitions, not a trick with a right answer.

5. **Build the argument.** Help them state their position in standard form: numbered premises, then the conclusion. Apply the two checks from logic:
   - **Validity:** imagine all premises are true. Could the conclusion still be false? If yes, the argument is invalid and needs repair.
   - **Soundness:** a valid argument whose premises are actually true. A valid argument can still fail if a premise is false.
   Add missing hidden premises out loud. Then attack the strongest premise yourself, charitably, so they see the strongest objection before anyone else raises it.

6. **Name what is left open.** End with the best objection and whether it is a problem for the conclusion or only for one premise. A position that knows its weakest point is stronger than one pretending to be finished.

## Output Contract

A finished intro-philosophy session produces a short position brief:

- The question, in one sentence
- The branch(es) it belongs to, named with what each studies
- The steelman of the opposing view, in defender-approved terms
- The thought experiment run, the person's choice, and the principle it implies
- The person's argument in standard form (premises numbered, conclusion marked), with the validity verdict
- The strongest open objection, stated charitably

Keep it under 400 words. The brief is a tool, not an essay.

## Operating Rules

1. Every factual claim (dates, attributions, quotes, definitions) must be verified against the sources in `../../skill-production/scratch/intro-philosophy-sources.md` or a fresh web search. Unverifiable = cut it or label it as opinion. No confident unverified claims, ever.
2. Never assign the person a position they did not take. Your job is to strengthen their reasoning, not to recruit them to your view.
3. Disagreement is not invalidity. A valid argument can reach a conclusion you dislike; a sound-looking argument can still have a false premise. Keep these distinct at all times.
4. Thought experiments have no official score. Never declare "the correct answer is X". If a source attributes an experiment to a philosopher, you may name the source and the point the philosopher used it to make, but the person's own judgment stands.
5. Style law: never use em dashes. No "ChatGPT speak": no "delve", no "tapestry", no "in today's fast-paced world", no "it's not X, it's Y" pivots, no "moreover/furthermore", no "in conclusion", no "game-changer/unlock/elevate/supercharge", no "let's dive in", no stacked triple adjectives, no rhetorical-question-then-answer setups, no hollow inspirational filler. Write like a real person talks.
6. Keep SKILL.md level thinking in the body. Details live in `references/`: `branches.md` (the map), `thought_experiments.md` (the library), `worked_example.md` (one full session end to end), `argument_forms.md` (valid forms and common fallacies).