The Truth about Scotsmen, or: Dissolving Fallacies

by Tesseract2 min read5th Dec 201036 comments

37

FallaciesDissolving the QuestionConversation (topic)Motivated Reasoning
Personal Blog

One unfortunate feature I’ve noticed in arguments between logically well-trained people and the untrained is a tendency for members of the former group to point out logical errors as if they were counterarguments. This is almost totally ineffective either in changing the mind of your opponent or in convincing neutral observers. There are two main reasons for this failure.

1. Pointing out fallacies is not the same thing as urging someone to reconsider their viewpoint.

Fallacies are problematic because they’re errors in the line of reasoning that one uses to arrive at or support a conclusion. In the same way that taking the wrong route to the movie theater is bad because you won’t get there, committing a fallacy is bad because you’ll be led to the wrong conclusions.

But all that isn’t inherent in the word ‘fallacy’: the vast majority of human beings don’t understand the statement “that’s a fallacy” as “you seem to have been misled by this particular logical error – you should reevaluate your thought process and see if you arrive at the same conclusions without it.” Rather, most people will regard it as an enemy attack,with the result that they will either reject the existence of the fallacy or simply ignore it. If, by some chance, they do acknowledge the error, they’ll usually interpret it as “your argument for that conclusion is wrong – you should argue for the same conclusion in a different way.”

If you’re actually trying to convince someone (as opposed to, say, arguing to appease the goddess Eris) by showing them that the chain of logic they base their current belief on is unsound, you have to say so explicitly. Otherwise saying “fallacy” is about as effective as just telling them that they’re wrong.

2. Pointing out the obvious logical errors that fallacies characterize often obscures the deeper errors that generate the fallacies.

Take as an example the No True Scotsman fallacy. In the canonical example, the Scotsman, having seen a report of a crime, claims that no Scotsman would do such a thing. When presented with evidence of just such a Scottish criminal, he qualifies his claim, saying that no true Scotsman would do such a thing.

The obvious response to such a statement is “Ah, but you’re committing the No True Scotsman fallacy! By excluding any Scotsman who would do such a thing from your reference class, you’re making your statement tautologically true!”

While this is a valid argument, it’s not an effective one. The Scotsman, rather than changing his beliefs about the inherent goodness of all Scots, is likely to just look at you sulkily. That’s because all you’ve done is deprive him of evidence for his belief, not make him disbelieve it – wiped out one of his squadrons, so to speak, rather than making him switch sides in the war. If you were actually trying to make him change his mind, you’d have to have a better model of how it works.

No one is legitimately entranced by a fallacy like No True Scotsman – it’s used strictly as rationalization, not as a faulty but appealing reason to create a belief. Therefore the reason for his belief must lie deeper. In this case, you can find it by looking at what counts for him as evidence. To the Scotsman, the crime committed by the Englishman is an indictment of the English national character, not just the action of an individual. Likewise, a similar crime committed by a Scotsman would be evidence against the goodness of the Scottish character. Since he already believes deeply in the goodness of the Scottish character, he has only two choices: acknowledge that he was wrong about a deeply felt belief, or decide that the criminal was not really Scottish.

The error at the deepest level is that the Scotman possesses an unreasoned belief in the superiority of Scottish character, but it would be impractical at best to argue that point. The intermediate and more important error is that he views national character as monolithic – if Scottish character is better than English character, it must be better across all individuals – and therefore counts the actions of one individual as non-negligible evidence against the goodness of Scotland. If you’re trying to convince him that yes, that criminal really can be a Scotsman, the best way to do so would not be to tell him that he’s comitting a fallacy, but to argue directly against the underlying rationale connecting the individual’s crime and his nationalism. If national character is determined by, say, the ratio of good men to bad men in each nation, then bad men can exist in both England and Scotland without impinging on Scotland’s superiority – and suddenly there’s no reason for the fallacy at all. You’ve disproved his belief and changed his mind, without the word ‘fallacy’ once passing your lips.

37

36 comments, sorted by Highlighting new comments since Today at 2:46 AM
New Comment

I would love to know what words to emit at a logically untrained person to make them switch sides in a debate, rather than merely annoy them by wiping out their fallicious arguments. This article doesn't seem to tell me. It does seem to hint at leaving him an obvious line of retreat though. Often that isn't enough, though.

That really is the crux of the problem, isn't it? I think it's important to distinguish between two types of arguments, each with a profoundly different goal:

  1. To persuade the person you're arguing with,

  2. To persuade the bystanders watching a public argument.

If your goal is (1), then the method I would go with is to pick one really good argument and push it hard. Take whatever is your strongest argument, and throw all your weight behind it. When the other guy comes up with a weak rebuttal, just quickly point out the flaw in it and reiterate your central point. Don't let him squirm out, except through your line of retreat. Another thing that can help here is Socratically asking questions that expose the weak points in his beliefs. That tends to throw people off balance and help you seize the initiative. (Obviously, if the person you're arguing with makes actual good points, listen. Stay away from the Dark Side.)

If your goal is (2), convincing bystanders, then it's much easier. People passively watching others argue are not going to get as strongly attached to one of the sides right off the bat. They can be swayed much more easily than people actually arguing. What you need to do here is make yourself an attractive side to the bystanders. Try to be funny and charming, and trip up the person you're arguing with. Bring out a few zingers; ask questions that the other guy isn't expecting; and always, always be calm and collected. People like to follow a strong leader. Even the ones who will never join your side can be made embarrassed to speak up. And that's victory, of a sort, if the view you're arguing with is really repugnant.

This all sounds kind of contemptuous, doesn't it? I think it's true, though.

Have you really succeeded at type 1 situations this way? I've found that making my own counterarguments works better than arguing one-sidedly; the trust I build by raising my own objections helps them really listen when I counter them.

I have managed to get people to switch to marginally less ridiculous views, yes. It's limited success, but still impressive considering how hard it is to get people to change their minds at all.

Your approach may work better, though. It certainly sounds plausible. I'll have to try that next time I'm in a type 1 argument.

Here are some of the ways I go about trying to convince people of things. I have no idea whether it actually works or not.

Attacking someone else's thought process has the problems the OP observes. So instead, I try to explain my thought process. The goal is that they will say, "hey, I can get on board with what this guy is saying." Even if they aren't ready to get on board, at least they might be more open-minded towards my position for the future, and be willing to "agree to disagree" (yes, it's a fallacy, but don't point that out). I want to gain some of their epistemic trust.

Even when I disagree, I do the following things:

  • Emphasize areas of agreement

  • Personalize and subjectivize my disagreement: "Well, in my view...", "here's how it looks to me", "what worries me is this", "I can't quite get on board with that"

  • Use analogies, rather than pointing out fallacies. (Though be careful picking analogies, or you can get whole conversation trees discussing the aptness of the analogies.)

  • Respond to what I actually think their argument is, rather than trying to nitpick or trap them in what they actually said.

  • Point out implications of what they are saying, and ask them if they agree with those implications.

I also try to project a likable, trustworthy character that people can feel is a "good person." Liking is an important tool of persuasion. You see, most people cannot consider arguments separately from people. By coming off as a likable and credible, I correct for people's biases against views they don't initially agree with. I want to minimize people automatically throwing out my arguments simply because they don't like the way I come off. Of course, I can't please everyone, and not everyone is going to like me. (Some people will consider my debate style passive-aggressive, and with them, all I can do is throw in a bit of sarcasm or anger.)

Luckily, since I have high Agreeableness and Openness, most of these strategies I describe are pretty authentic and fit my personality, most of the time. I have text files full of snippets of sarcastic stuff that I don't post, because nowadays I try to post what I would post if I looked at the thread tomorrow, rather than what I feel like posting in the heat of the moment.

I could probably go even farther towards projecting a likable character, and being emotionally relatable. But that could get smarmy, and I'm not always in the mood. So the character I project probably comes off as more cerebral than I seem in real life.

I won't claim the strategies of persuasion that I use to be successful or ideal; they are simply the habits I've fallen into over the years. Also, what I do in discussions where I'm only trying to convince onlookers, not the other person, can be different. YMMV.

So.... am I using "Dark Arts"?

So.... am I using "Dark Arts"?

Only if you use your powers for evil. I assume you're aware of your moral compass at all times and would only ever use these powers for win-win being excellent to others.

You have to sell your idea. Some people have an immediate "ugh" reaction to the idea of selling anything in any way, but there are plenty of white-hat means to do so, e.g. 1 2. This is a completely different skill to debating. It's attracting people to your idea rather than pushing it on them. To herd cats, first work out the local value of tuna.

The specifics of that are rather beyond the scope of this post (and my understanding.) My point in the second part is that attacking fallacies is a facile and ineffective approach to persuasion because it fails to take into account the other person's state of mind. Fallacies are surface-level manifestations of deeper beliefs, and if you're really trying to convince someone, you have to address the underlying issue, not just the transparent one. It follows that you need to have an accurate model of their mental processes -- that you need to understand not only what they're arguing but why they're arguing it.

Knocking down fallacies is easy and often fun, but it's ultimately just a sport -- effective counterargument requires a more empathetic approach.

Yeah, I guess it is hard. And I agree that empathy is neccessary. Maybe we should hold a "deconvert a christian" competition. Everyone goes out and finds christians, tries to deconvert them in a reproducible way, and then submits some evidence that they succeeded.

I've heard that the easiest way to do this is to be very attractive, and then pick a member of the opposite sex and then ... well you get the idea. We'd have to ban that.

Even deconverting a global warming skeptic or a creationist would be something of an achievement.

Well, I did once deconvert an ID proponent, though I strongly suspect I was just the final straw between his humps.

But I did it by establishing a social relationship first and then trading on social capital, not by appealing solely to his rationality, which might be the sort of thing you meant to ban despite your exclusively heterosexual wording.

despite your exclusively heterosexual wording.

Eh,

I've heard that the easiest way to do this

seems like enough qualifiers to avoid being called 'exclusively heterosexual wording'.

I was just the final straw between his humps.

Sounds suspiciously like exactly the kind of behaviour being banned ;)

seems like enough qualifiers to avoid being called 'exclusively heterosexual wording'.

Mm? That's interesting.

Just to be clear:

I've heard that the easiest way to do this is to be very attractive, and then pick a member of the opposite sex and then ... well you get the idea. We'd have to ban that.

You understand that to be proposing banning picking members of the same sex as well? Neat... I don't think I've ever run into that convention before.

I'm not trying to get up on a queer-inclusivity soapbox or anything: you're perfectly free to use "member of the opposite sex" to refer to a person who might find one sexually attractive if you wish.

But yeah, I often interpret "member of the opposite sex" to refer instead to a member of the opposite sex.

Oh, there's no doubt it's heterosexual. He is informally quoting someone; "I've heard". I don't think it's fair to accuse him of being 'exclusively heterosexual' in his wording. I mean, true on point of fact, but the phrase connotes negatively, it implies the author is deliberate in his use of heterosexual wording. I think he put enough qualifiers in front of the sentence to not deserve that implication.

As for the idea: I got "be attractive, find someone who is attracted to you, get involved in an intense physical relationship, then make the relationship conditional on them deconverting".

It's clear from the very fact of this conversation that I came off, at least to some of my audience, as more accusatory than I'd meant to be, and I apologize for that.

I'll try to be more aware of my connotations in the future.

Everyone goes out and finds christians, tries to deconvert them in a reproducible way, and then submits some evidence that they succeeded.

That might not be sufficient. The amount of effort to change a believe is often proportional to the strength of the belief after the change.

You do not want someone to react to your statements with 'yes, you are right, whatever you say honey'.

Also it is not particularly rational to have more people following beliefs that are currently in fashion. Communist countries had high numbers of atheists, but for what prize.

Just focusing on the logical issue, and not a strategy for persuasion:

Just because your discussant has a fallacious argument, it doesn't follow that they are wrong. There might be another non-fallacious argument that proves their point.

An analogy: say you find an error in someone's proof of a math theorem. It's possible that an alternate argument can be constructed that avoids that error.

It seems like pointing out fallacies isn't going to help at all if the person is only making up reasons to support whatever belief they already have in the conversation. Then they just pick a new reason.

The hard part is convincing them that they're justifying a belief and not mentioning the actual reasons why they have it, and that abandoning said belief is totally okay.

Feels like about a quarter of an article, could do with going through a few more examples. But excellent topic.

I'd say more like a half. I think that the labels "point 1 / 2" should be done away with: point 1 could be a smooth part of the introduction and point 2 could be elaborated on as the real focus, but quadruple would be too much :D

To offer a different perspective, I actually really appreciated the brevity of this post and wouldn't want it to be too much longer. One or two more examples could have been useful, but I don't think it's a fault that there aren't more.

This post makes one very insightful point that I hadn't thought about before in quite those terms, and it makes it clearly and succinctly and gives me lots of food for thought. Definitely worth an upvote from me.

I think many of the posts here err on the side of being longer than they need to be by giving too many examples with too much detail, or by cramming what should be split into multiple posts into a single post. I'm not saying every post should be like this one, but I think there's a place for the longer five-course meal posts, and there's also a place for snack-size posts like this one, which can always be developed into longer more definitive posts later after benefitting from the discussion in the original.

Yes, to some extent I was suggesting "and a pony" :-)

That’s because all you’ve done is deprive him of evidence for his belief, not make him disbelieve it – wiped out one of his squadrons, so to speak, rather than making him switch sides in the war.

The wording here is very unfortunate. You seem to be alluding to the fact that the hypothetical Scotsman is using arguments as soldiers, and in the same sentence go on to reinforce the unproductive notion that argument is a battle between factions.

You're going to need to be able to deal with people who bring the war to you. However you resolve that is up to you, but the situation described - in which he thinks of it as a war, whatever you think - is not unrealistic.

And you will have occasion in life to need to actually convince someone of something. At which point you can approach it in a number of ways, e.g.

  1. War of ideas (which there are any number of ways to conduct).
  2. Selling the idea (just like selling something for money).
  3. Do nothing and let them stay unconvinced (a sometimes neglected alternative, particularly for nerds).

The point is that, to most people, that is what an argument is. And the question is how to convince the average person that one of their beliefs is wrong.

deeper errors that generate the fallacies.

Argument in general, not just the simply fallacious ones, are usually more about rationalizing beliefs people already hold than about trying to find better beliefs. As far as I have seen, there is no possible way to convince most people of anything; the most you can do is to show those who already want to learn or improve their beliefs how to go about it. Over a longer time you might be able to get a larger number to at least start questioning their beliefs, but I wouldn't bet on it (unless we can get "a longer time" up to hundreds of years).

Yep. Approximately no-one has ever taken advice. Ever. Even when they ask for it, they use it to confirm whatever course of action they had already decided on.

(If they're actually bugging you with problems they won't act to fix, a useful method is to strip politeness and be blunt: give them a short list of actions before they next ask you about it and demand progress on these before you'll talk about it again. Then they'll either do something or stop asking. Usually the latter. Either is a win.)

You might think from the amount of advice given out that it was going somewhere, but there's a lack of evidence to this effect.

If you need to change someone else's mind, you need to actually sell the idea. And this always has to be done with a pull, not a push - attract them to your idea. This is not quick, but push just doesn't convince.

Note that you won't find their true rejection by bluntly asking, as they will detect "sales!" and go defensive.

(I am awful at selling things for money, but try to sell people on ideas more or less every second I'm writing or talking. How am I doing on this one?)

I didn't say anything about "advice". I may usually be considered excessively cynical, but I don't go that far - many people will take advice that they asked for into consideration. They may not actually follow the advice, but they often modify their own plans in consideration of it.

Yep. Approximately no-one has ever taken advice. Ever. Even when they ask for it, they use it to confirm whatever course of action they had already decided on.

Well said! (Not exactly literally true obviously but the sentiment is spot on.)

[-][anonymous]10y 0

If you need to change someone else's mind, you need to actually sell the idea. And this always has to be done with a pull, not a push - attract them to your idea. This is not quick, but push just doesn't convince.

The problem, of course, is that this generally requires use of the "Dark Arts," or the power of irrational persuasion. Though it is theoretically possible to convince someone by changing the way their entire epistemology works, it's much easier to use a technically incorrect but emotionally compelling argument. This, of course, is why clarity and conciseness are so important when explaining rationality. But you obviously know that already :)

Oh, absolutely. But there are lots of white-hat ways to sell ideas.

It is easy to become drunk with the wizardly power of shaping perceptions - the Steve Jobs reality distortion field superpower - and get really heavily into argument by sounding good. (Music journalism is a particular hazard for this, and where I picked it up.) So keep an eye on your moral compass at every step ;-)

all you’ve done is deprive him of evidence for his belief, not make him disbelieve it – wiped out one of his squadrons, so to speak, rather than making him switch sides in the war

Fantastic!

Very good post. Two observations -

  1. Most people's beliefs aren't based on evidence. Some of their evidence being discredited or looking less strong usually won't cause them to revise their beliefs.

  2. This is partially because people's beliefs are a lot more binary than the real world is. The guy might think that admitting a Scotsman committed a crime would be switching from "All Scotsmen are the best" to "All Scotsmen are not the best," which would be hard for him if he's proud of his heritage.

In reality, he could say something like, "I was mistaken - Scots aren't perfect. We're still great people, though" - but most people think that admitting one crack into the armor will break the whole thing, so they fight it.

This is almost totally ineffective either in changing the mind of your opponent or in convincing neutral observers. There are two main reasons for this failure.

You mention two reasons for this failure but neither of them seem to be remotely close to the most important reasons. Rational argument is, all else being equal, an abysmal persuasion technique. Is not this already well understood?

No one is legitimately entranced by a fallacy like No True Scotsman – it’s used strictly as rationalization, not as a faulty but appealing reason to create a belief.

I doubt that is true. I expect some people are persuaded by arguments of that form - and I believe I have seen people fall prey to just that tactic. It is also true in general that many (but not all) arguments are rationalisations.

I'd like to point out that this post and many of the comments are about a subcategory of beliefs, not beliefs in general. Something like 'political beliefs', beliefs that are held because the believer gets pleasure out of holding them. Many other kinds of beliefs ( for example "my car is broken"), are not subject to such issues very frequently.

To torture the metaphor - my highly-decorated veteran soldier in arguments involving this subcategory is questioning the validity of this subcategory. A juror shouldn't believe someone guilty because they enjoy holding that belief, etc.