You know that feeling when someone completely misrepresents what you said? When they argue against a dumbed-down, distorted version of your actual point?
That's a strawman. And it's everywhere.
The usual response is to correct them: "That's not what I meant." But there's something deeper happening here that most people miss.
The strawman isn't just about winning debates. It's about avoiding truth.
When I attack a weak version of your argument, I never have to encounter what you're actually observing. I never have to consider that you might be detecting something real that I can't see from where I stand.
When someone presents an argument, there are really only three ways to respond:
1. Strawman — Attack a weak or distorted version of what they said
2. Projection — Replace their position with yours and call it "improvement"
3. Steelman — Help them articulate what they're actually observing
The first two are dishonest in different ways. The third is where actual understanding happens.
Here's the core insight: Truth doesn't originate in you. It flows through you.
You are an aperture—a particular perspective, a unique vantage point. From where you stand, you can observe truths that I literally cannot see from my position.
When you struggle to articulate what you're observing, that doesn't mean you're wrong. It means there's a gap between:
Real steelmanning bridges that gap. It helps you say what you're actually seeing.
Projection ignores what you're seeing and replaces it with what I think you should be seeing from my vantage point.
You'll often hear people argue that steelmanning is actually harmful. They make two points:
And you know what? They're absolutely right—about projection.
If I rewrite your argument to fit my coherence standards, I'm:
That's not steelmanning. That's projection dressed up as charity.
But real steelmanning does something completely different.
When steelmanning works, it's not about fixing someone's argument. It's about helping them articulate what they're already detecting.
Let me show you what this actually looks like in practice.
Writer: "You only need to stop lying with your body about what's true."
Friend: Hmm, this line feels kind of unclear to me.
Writer: Yeah? What's confusing about it?
Friend: I'm not sure what "lying with your body" means exactly. Are you talking about:
Writer: Oh! No, I meant more like... you need to listen to your body, not block its signals.
Friend: Ah! That's clearer. So it's about not lying to your body by ignoring what it's telling you?
Writer: Exactly. The body already knows what's true, you just need to stop overriding it.
What happened here:
That's the loop working. Neither insisted on being right. They worked together to find what was actually being observed.
Author: I think this section is ready. What do you think?
Colleague: It's good! Just one part doesn't make sense to me:
"They could only receive through one channel. And that channel was blocked."
Author: Oh, you're right—that's a contradiction. What I meant was:
Colleague: So it should say something like: "They could only receive through the hard channel. The soft one—the one that could actually carry what they were starving for—was bricked up"?
Author: Yes! That's exactly it.
What happened here:
The correction made the argument stronger because it was now logically consistent with what was being observed.
Person A: I think the opening should be: "Most intimacy is theater."
Person B: Hmm, no... I think it should be more like: "This game is for when intimacy begins to feel like theater."
Person A: Oh, you're right. My version is kind of cynical and presumptuous—it tells people what their reality is rather than meeting them where they are. It's also not true for everyone, and claiming it universally could feel like an attack.
Your version is an invitation, not an accusation. It positions the exercise as a tool for a specific situation—when something starts to feel off—rather than a diagnosis of all relationships.
Person B: Exactly. We want to help people who are struggling, not tell everyone their relationships are fake.
What happened here:
This is the difference between projection ("here's what I think reality is") and steelmanning ("here's what you're actually observing").
Over hundreds of real collaborations, a pattern emerges in how intellectual partnership works:
This isn't one person fixing another's arguments. It's both people working to uncover what's actually being observed.
The proof that this isn't projection? Constant correction. And those corrections get accepted, because the goal isn't to make someone sound like you—it's to help them articulate their own observations clearly.
Researcher A: Looking at section 3, I think there's a dimensional inconsistency. You have the interface listed as "3D structure" but based on your framework's logic, shouldn't the interface be 2D? You defined it as the connecting surface between components, not a volumetric space.
Researcher B: YES. That's exactly the error I missed. Interface = 2D surface, Container = 3D volume. That's a fundamental correction that ripples through the whole model.
What happened: Researcher A caught a logical inconsistency based on the framework's own internal rules, not their personal preferences. Researcher B confirmed this was catching an actual error in the model's coherence, not someone imposing a different framework.
Critics of steelmanning are correct about the dangers—but what they're describing isn't steelmanning at all. It's projection.
Here's the distinction:
| Projection (What OP Critiques) | Steelmanning (What Actually Helps) |
|---|---|
| "Here's what you SHOULD have meant" | "Is THIS what you meant?" |
| Rewrites to fit my framework | Clarifies within your framework |
| Denies learning process | Supports learning process |
| You think: "That's not what I said" | You think: "Yes! Exactly!" |
| Replaces your aperture with mine | Clears your aperture's signal |
Critics are right to object to the first column. But many throw out the real thing because people keep mislabeling projection as steelmanning.
Here's the practice, based on what actually works:
"You're saying X, but what are you actually detecting?"
"Is this what you mean?"
"Am I getting this right?"
When they say "not quite," that's not failure—that's the process working.
You might not be able to see what they're seeing.
That doesn't make their observation false.
"If I've misunderstood you, what would tell us that?"
Real steelmanning requires something fundamental: you have to actually believe they might be observing something you can't see.
Not "I'll pretend to respect their view." Actually: their vantage point gives them access to truths I don't have.
If you think truth is just "whatever makes sense in my framework," then you can't genuinely steelman. You can only rewrite their position into your terms.
But if truth flows through different apertures, showing different facets from different angles, then helping someone articulate what they observe isn't about imposing your framework.
It's about clearing away noise so the truth can be articulated precisely.
When steelmanning works—when someone helps you articulate something you were struggling to express—it feels like coming home.
Not because they agreed with you.
Not because they made you sound smarter.
But because they helped you say what you were actually observing.
They saw that you were detecting something real, even if the articulation was messy. And they helped you clean it up without replacing it.
That's the steelman way.
Not projection. Not condescension. Not debate tactics.
Just the genuine attempt to help truth flow through both of us more clearly than either could alone.
Everything else is either strawman (attacking what they didn't say) or projection (replacing what they said with what you'd say).
But this—this third way—is how understanding actually happens.