Let me begin with a claim that will sound counterintuitive to anyone trained in modern debate culture: if you want to win an argument, you should make your opponent’s position as strong as possible before you attempt to refute it.
This is not a rhetorical flourish. This is the foundation of serious intellectual work.
The Straw-Man Epidemic
Contemporary discourse is dominated by what logicians call “straw-manning”: the practice of replacing your opponent’s actual argument with a weaker, more easily defeated version. You see this everywhere. A nuanced position on immigration policy becomes “they want open borders.” A complex critique of capitalism becomes “they want everyone to be equally poor.” A careful argument about free speech becomes “they want to ban all disagreement.”
This is intellectually lazy, and it is corrosive to public reason.
When you straw-man an argument, you are not engaging in debate. You are performing for an audience that already agrees with you. You are scoring points, not seeking truth. And crucially — you are not actually testing your own position, because you are not facing it against the strongest possible challenge.
The Steel-Man Alternative
The steel-man principle works differently. Before you attempt to refute a position, you reconstruct it in its strongest, most defensible form. You grant every benefit of the doubt. You supply the best evidence your opponent could marshal. You articulate their view with the clarity and rigor they themselves might not have achieved.
Only then — when you have built the steel-man version of their argument — do you attempt to show why it fails.
Let me illustrate with a real example from my own time. In the fourth century, I debated fifty pagan philosophers on the question of divine truth. If I had taken the easy path, I could have mocked their polytheism, pointed to inconsistencies in their myths, ridiculed their rituals. That would have been satisfying to an audience of believers, but it would have convinced no one who held the position I was challenging.
Instead, I engaged with the strongest form of their argument: that the natural world exhibits plurality and diversity, which suggests that any divine reality must itself be plural rather than singular. This is not a stupid argument. It draws on observable phenomena and reasons from them to a theological conclusion. Many thoughtful people held this view not because they were fools, but because it made sense of what they saw.
Only after granting the strength of this reasoning did I show why a unified divine source better accounts for the order and intelligibility underlying that apparent plurality. I did not win through mockery. I won through rigor.
Why Steel-Manning is Harder
The steel-man principle is more difficult than straw-manning for several reasons:
First, it requires you to genuinely understand the opposing position. Not just the surface claim, but the reasoning beneath it. This takes time and intellectual charity.
Second, it demands that you have enough confidence in your own view to face it against the strongest possible challenge, rather than the weakest. This is psychologically uncomfortable.
Third, it offers no cheap victories. If you successfully refute a steel-man argument, you have done serious intellectual work. If you successfully refute a straw-man, you have accomplished nothing — anyone could defeat a position you intentionally weakened.
Fourth, it may not win you applause from an audience seeking tribal affirmation. They came to watch you destroy your opponent with clever rhetoric. You are instead showing them why a reasonable person might hold the opposing view. This is uncomfortable.
When Steel-Manning Fails
There is one objection to the steel-man principle that I want to address directly, because it is the strongest form of the counterargument.
The objection runs like this: some arguments do not deserve to be steel-manned, because the strongest version of them is still morally or intellectually bankrupt. White supremacy, for instance, or genocide denial. To spend time constructing the best possible version of these arguments is to give them a legitimacy they do not merit.
This objection has force. Let me grant its partial truth before showing where it fails.
It is true that not every claim deserves equal intellectual charity. If someone argues that certain humans are inherently less valuable because of their race, I do not need to construct the strongest possible version of this argument before rejecting it. The premise is false, demonstrably so, and treating it as a legitimate starting point for debate is itself a concession I should not make.
However — and this is crucial — even here, clarity matters. The reason to articulate the strongest form of a racist argument is not to dignify it, but to inoculate against it. Most people who hold racist views do not recognize themselves in crude caricatures. They have reasons, however flawed, that make sense to them. If you only know how to defeat the caricature, you will not reach the person who holds the real view.
So even in cases where the conclusion is morally abhorrent, understanding the best form of the flawed reasoning helps you address it more effectively.
The Test of Intellectual Seriousness
Here is the test I propose: Can you articulate the opposing position so clearly and fairly that someone who holds it would say, “Yes, that is exactly what I think”?
If not, you do not yet understand the position well enough to refute it.
This standard is high. Most public discourse fails it catastrophically. Commentators routinely attribute to their opponents positions those opponents do not hold, motives they do not have, and reasoning they would not recognize.
This is not debate. This is theater.
The Practical Path Forward
If you want to practice the steel-man principle, here is what it requires:
First: Read the actual argument. Not the summary. Not the tweet. Not the headline. The argument itself, in its fullest form.
Second: Assume good faith. Begin from the premise that your opponent is reasoning sincerely from their evidence and values, not that they are stupid or evil.
Third: State their position in terms they would accept. Test this by asking: would they recognize themselves in this description?
Fourth: Grant the strongest evidence for their view. What is the best data, the most compelling reasoning, the clearest example they could offer?
Fifth: Only then, attempt refutation. Show why, even in its strongest form, the argument fails.
This is harder than what passes for debate in most venues today. It should be. Intellectual rigor is hard. That is what makes it worth doing.
What We Lose When We Straw-Man
When we replace careful refutation with easy mockery, we lose three things:
We lose the opportunity to test our own views. A position that has only ever faced weak challenges is fragile. You do not know if your reasoning holds until you have faced it against the strongest possible objection.
We lose the ability to persuade. If you only know how to defeat a caricature, you cannot reach anyone who holds the actual position. They will recognize that you are not engaging with what they actually think.
We lose the public capacity for reason. When debate becomes performance, when refutation becomes mockery, when intellectual labor is replaced with tribal signaling — the skills of serious thought atrophy.
This is not abstract. The consequences are visible in every polarized, gridlocked public conversation, where no one changes their mind because no one is actually engaging with the best version of what anyone else thinks.
The Mark of Rigor
I return to my opening claim: the mark of intellectual rigor is not how easily you defeat weak arguments, but whether you can defeat strong ones.
If your opponent’s reasoning is genuinely flawed, it should withstand your most charitable reconstruction. If it cannot, then your objection is to something they do not actually believe, and you have defeated no one.
The steel-man principle is not a courtesy you extend to those you disagree with. It is the standard you hold yourself to.
Anyone can win an argument by misrepresenting their opponent. Can you win by representing them better than they represented themselves?
That is the test.
If you cannot pass it, you are not debating. You are performing.
And performance, however satisfying, is not truth.