Steelman Then Critique
Can I trust this author?
Most criticism targets arguments at their weakest. A reader spots a clumsy example, a poorly worded claim, or a missing qualification — and attacks there, leaving the core thesis unchallenged. This feels productive but reveals nothing: dismantling a weak version of an argument proves only that the weak version was weak. The real question is whether the argument holds up at its best — when every charitable interpretation is granted, every ambiguity resolved in the author's favor, and every fixable flaw repaired.
Steelman Then Critique forces that question by splitting analysis into two explicit phases. In the steelman phase, you reconstruct the argument at maximum strength: clarify vague claims into their strongest possible readings, supply evidence the author could have cited but did not, resolve ambiguities in the direction most favorable to the thesis, and repair minor logical gaps that a sympathetic editor would fix. The result is the best possible version of what the author is trying to say — not what they actually said, but what they would say if they had unlimited time and perfect clarity. In the critique phase, you turn on this strengthened version and ask what still does not hold. The weaknesses you find now are genuine — they survive the most charitable interpretation, meaning they are structural rather than superficial.
This matters because most arguments contain both fixable flaws and real vulnerabilities, and readers who cannot tell the difference waste effort on the former while missing the latter. An argument that fails only because of a sloppy example is fundamentally different from one that fails because its core premise is wrong — but both look the same to a reader who attacks the first weakness they find. Steelman Then Critique separates the two, directing analytical attention to the load-bearing failures that actually matter.
The result is a two-layer assessment: the strongest version of the argument paired with its surviving weaknesses. Some arguments turn out to be far more resilient than their original presentation suggests — the author's poor writing obscured a sound thesis. Others reveal the opposite: once surface flaws are repaired, a deeper structural problem becomes visible that the original clutter was hiding.
Use this when
- You find yourself immediately disagreeing with an argument and want to ensure your critique targets real weaknesses rather than surface flaws
- The argument contains obvious errors or poor examples that make it easy to dismiss, but you suspect the core thesis may be stronger than its presentation
- Two critics of the same argument reach different conclusions about what is wrong with it — steelmanning first creates a shared baseline for productive critique
- You want to evaluate whether an argument's weaknesses are fixable (poor presentation) or structural (fundamental flaws that persist at the argument's best)
See this lens in action
Capital in the Twenty-First Century
The book's sweeping thesis about wealth concentration (r > g) is supported by extensive historical data but contains surface-level presentation issues that obscure genuine structural weaknesses — making it ideal for demonstrating how steelmanning separates fixable flaws from real vulnerabilities.
Product launching soonExamples
Economics/Policy
Thomas Piketty's "Capital in the Twenty-First Century" argues that the rate of return on capital consistently exceeds economic growth (r > g), producing inevitable wealth concentration absent deliberate intervention. A Steelman Then Critique analysis first strengthens the argument: grant Piketty's most favorable data interpretations, extend his historical analysis to cover the strongest supporting periods, and resolve his measurement ambiguities toward the thesis. The steelmanned version is formidable — centuries of data across multiple countries consistently showing capital returns outpacing growth. The critique phase then identifies what survives: (1) the model treats capital as a homogeneous category, obscuring that different asset types (land, financial instruments, human capital) behave differently under the r > g dynamic, and (2) the proposed global wealth tax assumes international coordination that no historical precedent supports. These weaknesses are genuine — they persist even in the strongest version of the argument, revealing structural limitations the original data presentation was obscuring.
Philosophy/Ethics
Michael Sandel's "The Tyranny of Merit" argues that meritocracy — the principle that success should track talent and effort — has become corrosive to democratic culture by fostering hubris among winners and humiliation among losers. A Steelman Then Critique analysis first grants Sandel his strongest case: accept his evidence that credentialism has intensified, that social mobility has stagnated despite meritocratic rhetoric, and that political polarization correlates with educational sorting. The steelmanned argument is compelling — meritocratic language genuinely does allow winners to moralize their success. The critique phase then reveals what still fails: (1) Sandel never distinguishes between meritocracy as an ideal and meritocracy as currently practiced — his critique of the implementation is presented as a critique of the principle itself, and (2) his proposed alternative — a renewed politics of humility — lacks any mechanism for how humility would be institutionalized without the very credentialist structures he criticizes. These are structural weaknesses that no charitable interpretation can resolve.
Common misapplications
Steelmanning so generously that the original argument disappears. If you find yourself constructing an argument the author would not recognize as their own, you have crossed from charitable interpretation into invention. The steelmanned version should be the strongest reading of what the author actually argues, not a different argument you wish they had made. The test: would the author say 'yes, that is what I mean, only better stated'?
Skipping the steelman phase and proceeding directly to critique. If you find yourself listing weaknesses without first constructing the strongest version, you are performing standard criticism — which risks targeting surface flaws rather than structural ones. The two-phase structure is the method: removing either phase produces a different (and weaker) analytical tool.
Treating the steelmanned version as the author's actual argument. The strongest possible interpretation is a reconstruction — a tool for finding real weaknesses, not a claim about what the author intended. If you find yourself attributing the steelmanned version to the author in your critique, you are manufacturing a target rather than analyzing the original text.
Don't confuse with
- Devil's Advocate →
Steelman Then Critique strengthens the author's position first, then finds its remaining weaknesses — you build the best version of the case for before critiquing. Devil's Advocate constructs opposition against the author's position — you build the strongest case against from scratch. Steelman asks "what is still wrong even at its best?"; Devil's Advocate asks "what would the strongest opponent say?" Use Steelman Then Critique when you want to find genuine structural flaws. Use Devil's Advocate when you want to stress-test resilience against organized opposition.
- Structured Self-Critique →
Steelman Then Critique is something the reader applies to the argument — first strengthening it, then finding weaknesses. Structured Self-Critique evaluates whether the author examined their own reasoning — did they acknowledge limitations and engage with objections? One is an analytical tool the reader applies; the other measures the author's intellectual honesty. Use Steelman Then Critique when you want to stress-test the argument yourself. Use Structured Self-Critique when you want to assess the author's self-awareness.
When to use what
| Situation | Use | Why |
|---|---|---|
| You want to evaluate an argument at its absolute strongest before identifying genuine weaknesses | Steelman Then Critique | Steelman Then Critique builds the best version of the argument first, ensuring critique targets structural flaws rather than surface errors. |
| You want to construct the strongest possible opposing case against the author's position | Devil's Advocate | Devil's Advocate builds organized opposition from scratch, while Steelman Then Critique strengthens the original before critiquing. |
| You want to find the unstated premises the argument depends on rather than evaluate its strongest form | Assumption Audit | Assumption Audit extracts hidden logical dependencies, while Steelman Then Critique tests argument resilience at maximum strength. |
| You want to evaluate whether the evidence supports the claims rather than assess the argument's best possible form | Evidence Quality Assessment | Evidence Quality Assessment rates the support material directly, while Steelman Then Critique reconstructs the argument before evaluating. |
Analytical checklist
Academic origin
The practice of strengthening an argument before critiquing it has roots in the philosophical principle of charity, formalized by Donald Davidson and Neil Wilson in the mid-twentieth century as the requirement to interpret a speaker's claims in their most rational light before evaluating them. In argumentation theory, Douglas Walton distinguished between "straw man" attacks (weakening an argument before refuting it) and "iron man" or "steelman" reconstructions (strengthening before critiquing), arguing that only the latter produces genuine knowledge about an argument's merits. The concept gained broader adoption in the rationalist community, where writers associated with LessWrong and effective altruism formalized the steelman as a prerequisite for honest intellectual engagement. In legal practice, the tradition of moot court embodies the same principle: arguing both sides at maximum strength to find the genuine points of vulnerability. Steelman Then Critique synthesizes this tradition for content analysis: rather than arguing in a seminar or courtroom, you reconstruct the strongest possible reading of a written argument, then identify the weaknesses that survive even that charitable interpretation.