It’s not though. At least according to every description of steelmanning I read you’re supposed to replace an argument you encounter with the best possible argument. But the best possible argument might be different enough that arguments which address it don’t address the original argument. Which is my whole problem with the practice. If you only make minor improvements to what I say that makes the arguments responding to my claims identical to the arguments responding to the new claim then I have no problems. But people in practice steelman arguments in ways that change the responses too them. At some point, if you aren’t actually addressing my claims we aren’t actually having a conversation.
I think the main disconnect here is the replcaing the argument point. You are supposed to take the best version of the argument presented, not replace it with the best possible argument. Of course anyone doing the latter is going to often be too far off course to move the conversation along.
To me steelmanning is more of a better version of restating the others point in order to verify you understand their point. The best steelmanning is often proceeded by that. (do I understand the other parties meaning, and what is the strongest version of that meaning.)
Most often though I just find that steelmanning is just ignoring fallacies in otherwise decent efforts at conversational debate. In order to not devolve into back and forth "thats a fallacy!".
> But the best possible argument might be different enough that arguments which address it don’t address the original argument.
In which case the responsibility is on the steelmanner to demonstrate conclusively why the best possible argument differs from the original argument by discussing what makes the original argument weak. You can't just address it with 'it's weak.'
