Steelmanned Counterarguments

Input: $ARGUMENTS

Overview

The opposite of a strawman. For any position, construct the strongest possible counterarguments — not weak ones you can easily dismiss, but the ones that would actually convince a reasonable person. Then assess which genuinely threaten the original position. This is how you find real weaknesses before someone else does.

Steps

Step 1: State the Position Clearly

1. What is the position being tested?
2. What are its key claims?
3. What evidence or reasoning supports it?
4. What does it assume?
5. What does it predict?

Step 2: Identify All Attack Vectors

Systematically find where the position could be wrong:

1. Evidence attacks: The evidence is wrong, insufficient, or misinterpreted
2. Logic attacks: The reasoning contains a fallacy or gap
3. Assumption attacks: A key assumption is false
4. Scope attacks: The position doesn’t apply where it claims to
5. Alternative explanations: The same evidence supports a different conclusion
6. Consequence attacks: If true, it leads to absurd or unacceptable results
7. Empirical attacks: Real-world data contradicts the position

Step 3: Steelman Each Counterargument

For each attack vector that applies:

1. What would the SMARTEST opponent say?
2. Not what a random critic would say — what would an expert who disagrees say?
3. Build the strongest version:
   • Use the best available evidence
   • Use the most valid logical structure
   • Assume the most charitable interpretation of the counter-position
   • Address the strongest form of the original argument (don’t dodge)
4. State each steelmanned counterargument as: “The strongest objection is that [X], because [evidence/logic], which undermines [specific part of the position].”

Step 4: Rate Each Counterargument

Strength criteria:

• Devastating: If true, the position is wrong
• Strong: Forces major revision of the position
• Moderate: Weakens but doesn’t defeat the position
• Weak: Technical objection, doesn’t change the conclusion

Step 5: Assess Which Are Genuine Threats

For each strong+ counterargument:

1. Can the original position survive this objection? How?
2. Does answering it require revising the position?
3. Is the counterargument actually correct? (It’s steelmanned, but is it TRUE?)
4. If you can’t refute it, does the position need to change?

Step 6: Report

STEELMANNED COUNTERARGUMENTS:
Position tested: [statement]

Counterarguments by strength:

DEVASTATING:
- C[N]: [counterargument] — Threatens: [what] — Status: [refuted/stands/needs response]

STRONG:
- C[N]: [counterargument] — Threatens: [what] — Status: [refuted/stands/needs response]

MODERATE:
- C[N]: [counterargument] — Threatens: [what]

Position after testing: [survives intact / needs revision / fatally weakened]
Recommended revision: [if needed]

When to Use

• Before committing to a position publicly
• Before making a high-stakes decision
• When you suspect your position has weaknesses you haven’t found
• When preparing to defend a position

Verification

• Position stated clearly before attacking
• All attack vectors considered (not just obvious ones)
• Counterarguments are steelmanned (strongest version, not strawman)
• Each rated by genuine threat level
• Honest assessment of which survive
• Position updated if counterarguments warrant it