Tags
I hate the word ‘steelmanning.’
Steelmanning refers to arguing with the best possible version of someone’s argument, even if it’s not the one they presented. Put like that, it sounds really good! After all, we all think it’s important not to misrepresent people [citation needed]; trying to present the best possible version of someone’s argument is good.
The problem is that every time I’ve seen the word ‘steelman’ used, it’s referring to one of two things.
In the least obnoxious case, Alice misinterprets and strawmans Bob’s argument, and then presents the argument Bob actually made as a steelman. That kind of steelmanning can feel really frustrating and condescending: not only is this person strawmanning you, but they’re also acting like you’re an idiot and they’re so much better than you for being able to think of the argument you actually made. However, it’s not that bad, because at least Alice does eventually get around to arguing with something that Bob actually said, however self-deceptive and strawmanny she is in the process. The best way to eliminate this sort of steelmanning is just to ask “did you mean [insert steelman argument here]?”
In the most obnoxious case, Alice doesn’t actually understand Bob’s argument at all. Often, there are fundamental worldview differences: for instance, Bob might be a Marxist, while Alice is not only a liberal but does not realize that non-liberals exist at all. That sort of steelmanning can feel like looking at your beliefs distorted in a funhouse mirror: Bob plaintively cries “but I don’t actually believe in autonomous individuals making decisions uninfluenced by society!” as Alice continues “now, the strongest form of ‘exploitation,’ I think, is that sometimes workers aren’t in a good bargaining position compared to employers, which can be totally solved by a universal basic income…”
That form of steelmanning is actively harmful to epistemic charity and to careful thought. Instead of understanding that people believe things differently from you, you’re transforming everyone into stupider versions of yourself that don’t notice the implications of their own beliefs. In fact, this kind of steelmanning is a form of strawmanning.
You can say “but neither of those are actually steelmanning! Real steelmanning is being able to put other people’s viewpoints in words they themselves find more compelling than their own arguments!” However, that is an extraordinarily rare and difficult skill; even most people who do it once can’t do it consistently. Saying “to steelman position X…” should be interpreted the same way as saying “to express perfect loving kindness for all beings…” It’s certainly a nice ideal which people might want to approach, and some people even manage to pull it off sometimes, but it’s a bit arrogant to declare that you’re definitely doing it. Even when you think you are, you usually aren’t.
What are the alternatives to steelmanning?
First, seek to understand the actual viewpoints people you disagree with are actually advocating.
Second, seek out intelligent and well-informed advocates of viewpoints you disagree with. You don’t have to make up what your opponents believe! As it happens, you have many smart opponents!
Third, whenever possible, try to switch conversations from a debate focus to a collaborative truth-seeking focus. This is best in one-on-one conversations between people who trust each other. You should be able to say “hm, but I’ve just noticed this piece of evidence against my position” without the person talking to you jumping on you and saying “ha! This proves I’m right! You admit it! I win I win!”– and they’ll do you the same favor.
Thing of Things 
I was super excited to read this piece because (a) I LOVE steelmanning and consider it to be the second best thing about the rationalist community; and (b) given (a), I am of course delighted by the prospect of being exposed to a strong argument against steelmanning, and of course, I wasn’t disappointed.
Here’s my 3 cents:
1) I agree that it’s disrespectful to say “Jim, your argument is dumb – let me make a better one for you.” And I agree that making both sides of an argument in your own head is potentially sterile.
2) With that, said, that’s not what I personally intend to express when I say steelman, including the many times I’ve used in in your comment threads. 🙂 I generally mean two related points:
(a) In a discussions about, for example, how exceptions for abortion restrictions could be logically consistent or what Larry Summers might have meant in his comments about women in science, I sometime use steelman as “I am not saying I find this argument convincing, but it’s the best argument I see for the position.” That frees me up to test arguments for conclusions that might be offensive, which I find philosophically useful. I’d argue that many people might benefit from “OK, I find conclusion X personally offensive. But what’s the best argument I can imagine that it’s true?”
(b) Sometimes I’m not focused on what a particular person means, but on the underlying point. In that case, I want all of the arguments for “should it be hard to root my phone” or “is the obligation to retreat a good idea or not given my values.” If someone can offer me what they see as the best argument for a side, I’m interested even if they don’t ultimately find it convincing.
3) Related to 2(b), sometimes a steelman is philosophically interesting. “Why did people think Gallileo is wrong” and “What was the best argument that Galileo was wrong” are different questions, both interesting in their own way.
tl;dr: IMHO, “steelmanning” is not great if you’re interested in why a particular person believes something. However, it is actually pretty great to test one’s own preconceptions, and to collect strong arguments when you’re interested in the underlying question.
LikeLiked by 5 people
Thinking a little bit more about my Larry Summers link above, I’m going to defend steelmaking a person as well, at least in some cases. When I characterized my interpretation of Summers’ “I don’t think there’s sexism in Harvard’s science faculty hiring, but I hope I’m wrong,” I was constructing what I saw as the most defensible interpretation of those words.
That’s also the principle of charity, but I’ve never successfully invoked that principle – I usually get a response along the lines of “so and so is a bad person, so the principle of charity is unrealistic.”. By using the concept of steelmanning, I could separate the question of whether Summers was correct and whether he was offensive, and concentrate on “what’s the best position that is consistent with what he said?”
LikeLike
Steelmanning is about presenting the strongest form of an argument, regardless of whether the person who made the original argument would agree with it. It’s not about accurately representing your opponent, but getting at the truth about an issue. For example, there may be many poorly grounded cases for X, but they have something in them that if developed make a much better case for it, even though it could involve abandoning some of the premises used by those who made the original arguments.
As I recall, presenting the strongest form of your opponent’s actual argument is called ironmanning.
LikeLiked by 5 people
Then steelmanning is objectively a fraud. There is no sensible way to parse “strongest argument for a false conclusion.”
LikeLike
Whatever the conclusion, some arguments for it are stronger than others – that gets more right, that has fewer incorrect premises, and so on. The steelman, then, is what you consider to be the strongest argument for a conclusion, regardless of whether it’s ultimately correct. To construct an extreme example, compare “abortion is wrong because it depletes the mana we all need to live” to “abortion is wrong because it causes fetuses to suffer”.
Also, it’s possible to discover/construct a persuasive steelman when the normal arguments by the proponents of the position are unpersuasive.
LikeLiked by 5 people
Also, “fraud” implies the intent to deceive. Playing with an idea is not inherently deceptive.
LikeLike
You have a lot of work to do sorting out the differences between “correct” and “convincing” before you can even start to try to justify statements like “an argument is stronger if it has fewer incorrect premises.”
LikeLike
The most likely case is that false means “believed likely to be false, ” which opens things up considerably.
LikeLike
Due to uncertainty, the correctness of arguments is always less than one, and quantifiably so. Steelmanning is replacing an argument with a similar one that has a higher probability of being correct, even if the probability is still less than other arguments.
LikeLiked by 1 person
Beyond basic stuff, the comparative strength of two arguments depend more on their fit to the audience than anything intrinsic to their phrasing.
Leaving out the “for me” in “Steelmanning is about presenting the strongest form of an argument for me, regardless of whether the person who made the original argument would agree with it.” is…
Self-centered, a little?
Rather rude to the person who made the original argument?
Basically exactly the second kind of Steelmanning talked about in the article, “the most obnoxious case”?
Like I’m just gonna leave this here:
“””
That form of steelmanning is actively harmful to epistemic charity and to careful thought. Instead of understanding that people believe things differently from you, you’re transforming everyone into stupider versions of yourself that don’t notice the implications of their own beliefs. In fact, this kind of steelmanning is a form of strawmanning.
“””
LikeLike
Typically, the difference between the arguments isn’t just in how they’re tailored to their audience but in their actual content – premises, evidence, and so on.
We enter into debates under the presumption that we’re right and that whatever general worldview we happen to hold is correct. Thus, it’s fine to leave out the “for me”, because it’s not just about fitting arguments to people, but to what the speaker considers to be reality.
Finally, steelmanning doesn’t presume that other people are stupid, nor that they don’t have different beliefs from you. People don’t have to be stupid to be wrong, nor (and this is the heart of steelmanning) do they have to start with the same premises to come up with a worthwhile argument, even if it’s not great as presented.
LikeLike
The statements “the argument this person presented to me is not the strongest possible argument for the advocated position,” and “it is an act of arrogance to presume that a person who disagrees with a position can come up with, on their own, a better argument for said position then any person who actually agrees with it,” are consistent with each other. Even if you could come up with an argument that is more likely to convince you of a position then the argument of the person you are currently talking with, I agree with the apparent claim of Ozy’s that, if the position is one you currently disagree with, there exists an argument that is more likely to convince you of the position then any argument you could come up with yourself. Perhaps you should put forth some effort to find this argument before trying to formulate it yourself. I can appreciate the existence of situations in which one can look for and not find this argument (because of difficulty in being found not because of its non-existence) but still, it would usually be more worth your time (assuming you want to find the best argument for a position you currently disagree with) to look for such an argument first before trying to come up with one yourself.
There are also differences between behavior when actively engaged in a conversation, behavior when on one’s own trying to explore possible positions, and the reports of such things. Good behavior in one setting may be bad behavior in another. Steelmanning a position one doesn’t hold while engaged in a conversation with someone who does hold the position should be evaluated differently then steelmanning a position when one is by oneself exploring the position. If I gained insight on something by doing the later, I would report it as “the argument that I have found or been presented with that is most likely convince me,” rather then as “the best possible argument.” The former is not only more likely to me more correct but is also more polite.
In general I approve of the approach to argument that is “two (or more) people hold contradictory beliefs and an argument should be these people trying to, together, discover something which, for each person, is more likely to be true then the position held by that person at the start of the argument,” then the approach that is, “I should use argument to prove myself correct to people who currently disagree with me.” Not only is the former more likely to be successful then the later in achieving its goals but it is more likely to cause one to migrate one’s beliefs to those more likely to be true. I do not necessarily think that you disagree with this approach nor do I think it is the exact same as the approach advocated by Ozy but I think it is valuable and related. I also understand that not everyone will co-operate with the former approach but it is more likely someone will when approached with it then when approached with the later approach. I also appreciate that the two approaches are not always clearly differentiated from each other.
Finally, I want to re-iterate and emphasize that attempting to find the best argument for a position that already exists before attempting to formulate the best argument for a position oneself is usually a more productive path to finding the best argument for a position then the reverse is. I do think (but am not sure) that this is something advocated for by Ozy in this post.
LikeLiked by 1 person
We live in a much less convenient world than that.: there is no universal scale of argumentative goodness. As in the example for the marxist and the liberal, people can disagree about fundamentals, even about pistemological fuindamental like what constitutes a fact or law of inference. Which, incidentally, is why Aumann’s Agreement Theorem is rubbish.
LikeLike
One strange variant of this phenomenon that I’ve noticed is that whenever there’s yet another circlejerk amongst Rationalists about how much [outgroup, usually feminists] sucks and makes such terrible illogical arguments, someone will say something to the effect of “There’s a steelmannable point there, where…” as sort of the Rationalist-shibboleth translation of “To play Devil’s Advocate for a moment,” sort of a way to say, “I’m [outgroup] but not one of THOSE [outgroup member], I promise,” and evade the natural defensiveness that only people who think they’re immune to cognitive biases can truly possess.
LikeLiked by 2 people
There’s definitely some irony to a group putatively dedicated to addressing cognitive bias deciding that the best way to refute a conclusion is to give the job of summarizing and advancing that conclusion to someone who doesn’t agree with it, and is summarizing and advancing it for the specific purpose of defeating it in a show trial.
Can you imagine applying that elsewhere?
“Your honor, I move that the defendant not offer his own defense. I, as the prosecutor, will summarize the best defense he could have offered, then show it’s flaws. Trust me.”
LikeLiked by 1 person
There should be a rule to never steelman unless you have failed to strongman.
LikeLike
There’s no material difference between strawman and steelman arguments. In both cases a critic of the conclusion at issue invents a position that isn’t actually held by those he’s critiquing, creates it knowing he doesn’t believe it, creates it for the purpose of shadow boxing it to death in a demonstration match in which the actual position of his opponent has neither presence nor advocate, and then having destroyed his own creation declares victory over his foe’s actual position and conclusion.
LikeLiked by 2 people
Well, I mean, there is. Like, you aren’t granting the premise. If you think that all steelmanning is just strawmanning in disguise, then that’s fine, but that would be different from saying that trying to make people sound better and trying to make them sound worse are the same thing.
LikeLiked by 6 people
Firstly, I really like: “You don’t have to make up what your opponents believe! As it happens, you have many smart opponents!” A very good reminder, and another point in the “let’s not throw intelligence at everything and reinvent the wheel” bucket.
Second, these failure modes definitely make sense to me, and I’m glad you’re bringing them up. I guess, in addition to not coming off as condescending, it’s important to distinguish between:
1. Restating an argument in a way your interlocutor would agree with
2. Making their argument better by adding information and ideas you have and they don’t. Like “I’m not sure if i agree, but you’re right that X, in fact I just read a study that sports X because Y”
3. Making the best argument for a position from their perspective and premises, sort of from scratch, often to a third party
4. Basically passing an ideological Turing test of your conversational partner’s position (probably overlaps with one or all f the above)
5. Making the best argument for a position from your perspective and premises
And which you do will depend on your goals and relationships!
(Is one of these maybe charity rather than steel manning?)
Good faith truth seeking uber allies, for sure.
LikeLiked by 3 people
I agree!
The other thing that seems important to me is that a lot of times ‘steelmanning’ seems to be coming from an adversarial framework (this is kind of an unfortunate hangover from analogy to ‘strawman,’ I think– although I notice you’re careful to say ‘interlocutor’, which is a more neutral term, and that’s good). I think that it’s a really good idea to say evidence you see for things you disagree with, but IME it’s easier in a collaborative discussion than in an adversarial one. So I think the way I prefer to frame it is less about the specific tactic of steelmanning and more about the whole way you approach conversations? OTOH, your original post was probably not meant to be a complete exploration of good discussion norms. 😛
LikeLiked by 2 people
Thank you for presenting these objections to steelmanning. They’re interesting and I hadn’t thought of them. I agree with objection one, when you’re steelmanning a person you’re not really being respectful to them.
I have a lot of sympathy for objection too (as a Marxist myself, your example certainly made me laugh). I do think this got a bit too theoretical though, and could benefit to be more grounded in examples we’ve experienced.
It’s been my experience that I read a terrible editorial (like https://www.theguardian.com/commentisfree/2014/apr/08/beware-of-cupcake-fascism ) and think “well the author is rude and arrogant and has a bunch of bad examples – but I should not ignore the core point they have.” Sometimes I can even convince myself the core point is right. So it’s helpful to pare down the arguments from the separable parts you find annoying, and build it up some in the terminology you would use, and see how you feel about the central idea. That seems a useful skill to have (though yes, done badly this can mean you rephrase everything through your own myopic ideology.) Especially in cases where you can’t dialogue with the author, I’m not sure what a better option is.
I’m interested in hearing your examples of when steelmanning has gone wrong.
I really like your collaborative truth-seeking focus suggestion, and will generally try to lean towards that. It may not be any more humanly achievable than your other examples of ideal interaction earlier on though. Something to strive for, but good to have coping strategies when it fails.
LikeLiked by 1 person
I’ve had the unfortunate experience of someone demanding I ignore an egregious insult because I should be steelmanning my opponent’s position, implying that since the steelmanned position wouldn’t have such a horrible insult I should allow the actual insult to pass without comment.
LikeLiked by 3 people
Was this a case of a position that you found insulting, or an unconnected insult tacked onto some position?
LikeLike
A case where, in order to “strengthen” their argument, a person took the maximally offensive interpretation technically allowed by the literal meaning of the words others said. More specifically, “believing that there are statistically significant population level differences in ability and/or preference between races and sexes” was equated to “supporting segregation” and “wanting to treat blacks as second class citizens”.
A third party took issue with the response this garnered, that people were arguing with what was actually said and not the sanitized, steelmanned version. That people weren’t being gentle enough because they were trying to “crush” bad arguments rather than acknowledge things that weren’t actually said.
LikeLiked by 1 person
Oh, and “connected to ethnic cleansing and all kinds of other brutalities”.
That’s right, thinking that ashkenazi jews are overrepresented among nobel prize winners because of population level genetics means you support racial genocide! Fuck statistics, we have the non-central fallacy![/sarcasm]
LikeLiked by 1 person
Well, don’t argue with idiots. They will drag you down to their level and beat you with experience 🙂
LikeLiked by 1 person
I often see *accusations* of steelmanning, which fall into the nasty pattern of “one person presents a reasonable argument, another person engages with it, and a third person, who has already mentally strawmanned the first person, accuses the second of steelmanning”.
LikeLike
It is extremely unlikely you can steelman a position you fundamentally disagree with. However sometimes you can “steelman” a philosophy you have some disagreements with. The best example I can think of is Scott’s “Neoreactionary philosophy in a giant planet sized nutshell.” Scott has many disagreements with neoroeaction. But he also has many agreements and clearly thinks neoreaciton has produced many important insights. One could think of Scott’s Nutshell and a summary and “Translation” of neoreaction aimed at Scott’s intellectual sub-culture.
However I find it extremely unlikely Ozy/Topher/etc could steelman neoreaction. And I find it a bit unlikely Scott could steelman feminism. Though possibly any of them could make “steelman” arguments for specific positions of feminism/neoreaction.
Its also worth noting that while Scott does not fully endorse neoreaciton he is NOT trying to dis-credit them. Scott lieks their work and while he won’t admit it he is a major source of publicity for the movement (which he wisely refuses to comment on).
LikeLike
I think I can steelman positions I disagree with, given the caveats that (1) I’m not smart enough to likely to come up with the literally “strongest” argument for anything; and (2) my arguments tend to be centered on things I find important. (Well, and (3), I often do so by reviewing a bunch of other people’s arguments and synthesizing them).
But with that said, I find that I can often get to an argument that breaks down to a couple disputed priors. I think I know what I consider to be strong candidates for the strongest arguments for and against making cake bakers cater gay marriages, or for and against whether George Zimmerman is guilty of murder, and for and against trade liberalization, for example.
LikeLike
I feel that this misses one of the key points of a good steelman. If there’s one thing that can be said unambiguously about rationalists, it’s that we’re familiar with our own first resort counterarguments. That means that, when constructing a steelman intended for other rationalists, we can preempt the first eight or nine “obvious” responses and save about half an hour.
LikeLiked by 1 person
The chain rule state that, if F(x) = f(g(x)), then F'(x_0) = g'(x_0) f'(g(x_0)) (assuming f and g are each differentiable in the appropriate location.)
Recall, the derivative of f at x_0 is lim h -> 0 ((f(x_0 + h) – f(x_0)) / h). Thus, defining F as above, we wish to find lim h -> 0 ((f(g(x_0 + h)) – f(g(x_0))) / h). Multiplying by ((g(x_0 + h) – g(x_0)) / ((g(x_0 + h) – g(x_0)), we get lim h -> 0 ((f(g(x_0 + h)) – f(g(x_0))) / ((g(x_0 + h) – g(x_0)) + lim h -> 0 ((g(x_0 + h) – g(x_0)) / h, which gives the correct result, since we’re presuming f and g are differentiable at x_0.
This argument is invalid, although the conclusion is correct. In particular, when we multiply, there’s no guarantee that the denominator is nonzero in a sufficiently small neighborhood around x_0 (eg g is constant on some interval around x_0.)
However! This invalid argument contains some good ideas. Enough for a mathematician to come up with a clever way of using the same basic idea to salvage a valid proof.
If you come from a tribe that believes the chain rule, you probably won’t bother checking the argument that looks valid, even though it’s not. If you’re from a tribe that doesn’t believe the chain rule, you will probably know very well why that argument is wrong, and feel smugly superior to those chain-rulers who don’t understand how limits interact with dividing by zero.
But, if I’m interested in believing true things regardless of tribal affiliation, the clever mathematician who can correct the flaw in the chain-ruler proof is valuable to me, even if this is different from the arguments made by every chain-ruler.
I am extremely confident that most people don’t know the strongest arguments for most of the things they believe. And why should they? Even if their beliefs weren’t mostly a result of who they grew up with and only went with the beliefs that had the strongest arguments, unless they make a heroic effort to not have opinions until they had thoroughly investigated the strongest arguments for every side available (which are hard to find courtesy of Sturgeon’s law, even presuming that every side has actually figured out their strongest arguments), then their beliefs will reflect the strongest argument they’ve heard, which is quite different from the strongest argument articulated (which is, again, different from the strongest argument articulatable.)
LikeLiked by 3 people
My post is not about whether it is theoretically possible to come up with better arguments for someone’s position than they have; after all, I said that that does happen sometimes. My post is about whether instances of that happening are more than about one percent of alleged steelmans and whether the majority of the rest are just excuses to strawman people.
I agree that most people don’t know the strongest arguments for things they believe. Why are you talking to most people? As I explicitly stated in the post, find the smartest people you disagree with and see the arguments they make.
LikeLiked by 2 people
To challenge you a little, Ozy, I googled the use of steelman in your comments, and I didn’t personally think that any were an excuse to strawman the arguments. Especially mine of course, but also everyone else’s. 🙂
LikeLike
I’m also reminded of that post where someone on Tumblr said (paraphrasing from memory) that the “Principle of Charity” should be renamed “The Principle of being nice to my ingroup and smiling through my teeth at my outgroup while passive-aggressively sniping at them until some of them snap at my bait and thus prove that my hatred for them is rational because of their totally-preemptive hostility towards me.”
What better way to do THAT than to loudly proclaim how I like to “Steelman” my outgroup’s arguments because gosh darn it they’re just too irrational to make a good case themselves and so I have to do it for them, the poor delusional darlings…
LikeLiked by 1 person
OK, here’s two uses of mine.
1) Patrick (who I respect and enjoy talking to) argued that Michelle Bachman was an idiot for saying the phrase “separation of church and state” isn’t in the constitution. I argued that if you steelmanned her statement (a) it’s literally true, and (b) can be construed as an argument that while the establishment of a state religion would clearly violate the constitution, reasonable people could argue that the courts have gone too far in blocking arguably trivial overlap like creches on public property, or voluntary non-denominational school prayer.
As it happens, I don’t agree with that argument, but I guess I think it’s within the realm of reason. IMHO, steelmanning is helpful for the question of “does pointing out that separation isn’t in the Constitution establish that she’s an idiot, or does it make a point that’s within reasonable discourse?” (To be fair, Patrick argued that her other work forecloses any rational interpretation, and I’m not familiar enough to judge that).
https://thingofthings.wordpress.com/2015/01/21/people-are-people-regardless-of-political-affiliation/#comment-5344
2) Nita asked for a “charitable interpretation” of Larry Summers’ famous comment about women in science (that he didn’t think it was because of discrimination, but he hoped he was wrong). I offered what I characterized as a steelman version of his argument, which was the best argument I could come up with that was consistent with what he said.
https://thingofthings.wordpress.com/2015/01/27/open-thread-7-war-plans/#comment-5836
Maybe I was using steelman wrong – it was what I guessed Summers actually did mean, but I was consciously trying to put the argument in a more developed and articulate way than he did.
I might be using steelman all wrong (I don’t think so), but in any case, I’m as confident as it’s possible to be that I don’t use the term either with the motives or the effect that you describe.
It might not be kind to call people out, so maybe we shouldn’t go into cases, but I submit that none or almost none of the uses of “steelman” by Ozy or xir commenters meet your description
LikeLike
Well, I suppose that means I wasn’t talking about you, then…
LikeLiked by 1 person
IMHO, steelmanning is only a useful concept inside your own head. If you have to claim to be steelmanning a position, you are very probably doing it wrong*.
In its original sense, I mostly understood it as a reminder not to pounce on technical errors in an argument if you know what the other person meant and could easily fix the argument. You would not write about it, you’d just do the substitution in your head and instead attack the points you don’t think are true or fixable.
Attacking trivial and easily fixed technical errors is a cheap rhetorical trick, and only wastes everyone’s time (unless you have to win a debate using Dark Arts).
*or using the term as a rationalist-flavored version of devil’s advocate. But then, you should probably just say you’re playing devil’s advocate.
LikeLike
So I would call myself “almost rationalist adjacent.” I discovered blogs like this one, Alas, SSC, and several very entertaining tumbler blogs in my mid 30’s. I’m a blue collar guy (welder by trade) with no college degree, and not so well-read as most of you. I have time-consuming sports and hobbies that I love, and not much time for books, sadly. I haven’t had time to read through much of the sequences, and I don’t have the base knowledge to follow along with everything (maybe most things?) I read on LW. I rarely comment here or elsewhere because I have little to contribute. I gobbled up every article and comment here, though.
To me, steel-manning shouldn’t be used in discussions so much as it should be used inside one’s own head to help guide the way they discuss things with others. To you guys, being charitable is probably a habit by now, but in my world almost no one bothers arguing against their opponent’s best arguments.
Rationalists discussing things among themselves have no need for steelmanning. Most of you are in the habit of always being charitable, and always being informed on both sides of the most common debates found on blogs like this one. I don’t expect to read about some great new idea or insight from somebody who is performing a steelman.
I do think that getting in the habit of steelmanning will help those who aren’t familiar with rationalism, people like me a couple of years ago. I would be delighted if more people in my life were familiar with the concept. I have educated friends all around me, but no one to have a stimulating discussion with, and it’s lonely in a way.
LikeLiked by 2 people
Keep posting and participating! I’m trying to do that more too, and to get myself a little less lonesome, both intellectually and emotionally.
Also, it really seems like steel manning has both a light and dark arts version. Any potentially useful technique for making an argument more productive could also be hollowed out and used as a sock puppet by the parasite that is cognitive bias, or by conscious malice. All our brains have those two worms burrowed into them to some extent, but I don’t think that means literally every person in every instance is using steelmanning or other techniques for manipulation rather than truth-seeking, even if it could well be the majority of instances.
LikeLike
Also, I like the vibe of having these sorts of concepts accessible and useful as ways of clarifying thought for people who aren’t already massively privileged in terms of education and their general intellectual life.
LikeLike
I wonder if steelmanning may be thought of as an ideological translation (similar to the ideological Turing Test). When a Marxist and a liberal disagree, sometimes it takes a Marxist to translate the liberal’s arguments into a form that fits into the Marxist paradigm without changing its essence.
LikeLike
In your examples, Alice is in conversation with Bob and trying to steelman his arguments. I’m all for discouraging this in favor of listening to and understanding Bob. This surprises me, because I usually hear steelmanning invoked in a different context. Suppose Alice and Bob are discussing a belief held by Charlie who is not present. Both Alice and Bob disagree with Charlie’s conclusion and have not been convinced by what they know of his arguments. Before rejecting the conclusion outright, they try to steelman the argument, constructing the most compelling case they can make for Charlie’s belief.
What do you think of this kind of steelmanning? Charlie’s absence seems to avoid the rudeness and much of the condescension concerns. However, it remains true that Alice and Bob may not properly understand Charlie’s argument or his conclusion and probably won’t succeed in constructing the best possible case for it.
LikeLike
To me steelmanning is to be done AFTER youve finished looking at, and have already rejected, the argument as stated. Its basically ‘im not convinced by this as it is, but lets not dismiss it completely until ive tried as hard as I can to get anything of use out if it.’.
Its to be done by yourself or when the people eho made the original argument are not present. Its basically playong devils advocate (which you dont tend to do in an argument with the devil themselves). Also I dont know how playing devils advocate seems to now have a negative connotation, but I lile devils advocates and think ots a useful thing to do.
LikeLike
Pingback: The Steel man Undisclosed – Radical Reason
Pingback: Against against steelmanning – Thomas's notebook
I’ll confess to not fully understanding the point of this post. Is it about the incorrect application of the “steelmanning” strategy? Or is it claiming that correct “steelmanning” is impractical / intractable? Your 3 steps at the end look almost identical to Daniel Dennett’s 4-step debate process, which has been called a “steelmanning” approach, so wouldn’t that mean that you are somehow advocating for a more correct way to “steelman” than the two (obnoxious) examples you gave?
My interpretation of this post is that you created an anecdotal strawman against “steelmanning” in order to advocate for… a better kind of “steelmanning”?
LikeLiked by 1 person
Pingback: Rationality Recap: Steel Man – Vincit Veritas
Pingback: Review of Ernst Cassirer: An Essay on Man (1944) – Homemade multibody dynamics