In spaces where people deliberately learn critical thinking, it’s common for people to learn a list of logical fallacies–in much the same way that one might learn a list of names of Pokémon. Then, to reinforce this knowledge, we start spotting the fallacies in the wild. It’s a good learning exercise. But once all is said and done, and you’ve successfully internalized the list, what then? Is fallacy-spotting a good way of engaging with arguments?
I don’t think so. I’ve long said that it’s obnoxious and unproductive to explicitly name fallacies in the course of an argument. But even if you keep it to yourself, I think fallacies are an extremely limited and misleading way of engaging with arguments.
Ad hominem: a case study
“Ad hominem” is a logical fallacy where you attack a person instead of their argument. Suppose that we have someone guy named Joe, and he argues that cars are bad for society. He cites pollution, accidents, inefficient commutes, and so on. You agree with Joe, but I disagree. (IRL, I agree with Joe, but this is just a hypothetical for illustration purposes.) But rather than discussing anything about cars, I talk about how Joe is a bad person. I’m engaging in ad hominem.
However, there are many variations that look like ad hominem, but don’t quite fit. For example, suppose that I attack Joe’s character and I address his arguments. What is going on there? Several possibilities:
- Attacking Joe’s character is just window dressing, in the same way that I might begin with a personal anecdote to engage an audience.
- I am making two distinct points: that Joe is wrong, and that Joe is a bad person.
- You believed Joe in the first place not merely because of his words, but based on an assumption of wisdom, trust, and good character behind them. When I attack Joe’s character, I am trying to correct the bias that you had in his favor.
We could go deeper on each of these possibilities. For instance, if it’s window dressing, we could question the wisdom of its inclusion. Is it just a sneaky way of engaging in ad hominem arguments, but with plausible deniability? Or perhaps the problem has nothing to do with the argument itself, but rather that attacking people is socially inapropriate, or discourages us from resolving arguments. If I have made inappropriate attacks on Joe, what do you do about it? What “punishment” is appropriate?
Possibility #3 suggests that an attack on Joe’s character could function as a legitimate stand-alone argument. Joe claims a bunch of things about cars, and you believed him. But I argue that Joe is untrustworthy. I don’t have to say anything about cars, I just have to put you back at square zero. This can be a legitimate argument, and that forces us to re-examine even the straightforward examples of ad hominem.
I use ad hominem as an example, because in these parts, I think it’s generally agreed that ad hominem has a lot of issues going on with it. For example, in a recent thread on Pharyngula, someone accuses PZ of ad hominem, and PZ blows it off. And the commenters all chime in to talk about what is and isn’t an ad hominem, including someone linking to a 300 page dissertation on the subject.
My claim is that ad hominem is not special. Every single fallacy is like that. Every last one of them is subject to complex variations, loopholes, and social implications. It’s good to be familiar with logical fallacies, but if you approach an argument with the mindset of spotting them, you’re missing so much.
Logical arguments vs narratives
Most arguments are not about logical inferences. Arguments are about warring narratives. I tell a story, you tell a story, and we each decide what story is most compelling. That’s how I’m arguing now, it’s how you would argue if you wanted to rebut me. That’s just the reality of the situation.
The problem is that logically valid arguments severely underdetermine the truth. Perhaps relying on stories is bad, and it would be better to stick to logically valid arguments (and I mean “logically valid” in the broad sense to include induction and probabilistic arguments). But logically valid arguments are insufficient to address most questions that are important to us.
In the context of warring narratives, fallacy-spotting leads to applying an arbitrarily higher standard, basically whenever we feel like it. When engaging with an argument that we dislike, we put our fallacy-spotting into high gear, and find reasons to reject it. And when arguing for our own viewpoint, we engage in the same storytelling as everyone else. We would become entrenched in the vast majority of our beliefs, since nobody could mount a logically valid argument against them.
One way of thinking about a logical fallacy, is that it’s a narrative that initially seems compelling. You counter a logical fallacy with an alternative narrative. For example, ad hominem tells the narrative of a person who is bad and wrong. To counter ad hominem, we propose the alternate narrative, that someone can be right regardless of whether they’re a good or bad person. Logical fallacies are a good storytelling toolset, keeping us aware of all narrative possibilities. But eventually we get so good at storytelling, that we can tell any story we like.
Course correction
When we encounter a logical fallacy, is the appropriate response to nullify the fallacy, or to punish it? Do we act like the fallacious argument was never made to begin with, or do we try to counter the fallacy by mentally biasing ourselves in the opposite direction?
In my ad hominem example, I attacked Joe’s character. So you could make a course correction by loudly naming the fallacy and rejecting the argument. You refuse to consider the rest of my argument until I withdraw the ad hominem. You tell everyone to stop listening to me because I’m a bad person who makes ad hominem arguments.
Formally speaking, it’s incorrect to punish a fallacy (this is known as the “fallacy fallacy”). But if you don’t punish fallacies, doesn’t that permit me to use them with impunity? Even if you recognize my argument as an irrelevant ad hominem, it may have still had a subconscious impact on your views. It may have impacted the views of people in the audience. Or perhaps you only spotted half of my fallacies, while the other half are still in force.
I’m fascinated by the idea that some fallacies are epistemological course correction. For instance, you placed your trust in Joe, and believed what he said. In attempt to course correct, I attack Joe’s character. Then you try to course correct by accusing me of ad hominem. I course correct again by accusing you of the fallacy fallacy. It’s like we’re fighting over the wheel of a car.
What are we arguing about?
Another common theme of logical fallacies is disagreement in subject matter. For example, attacking a person’s character is not ad hominem if the person’s character is precisely the subject in question. “Begging the question” is essentially a mistake in subject matter, where one person assumes what the other person believes is the object of contention. A “straw man” is when I argue against a particular viewpoint, and you think I should have argued against a different viewpoint.
There is obviously no way to “prove” that the subject matter is one thing or another. The subject matter is, more or less, whatever we choose it to be. And if we disagree on what the subject matter ought to be, we have to negotatiate goals. For instance, is your goal to persuade me? Then you cannot persuade me without starting from what I believe. You cannot beg questions or use straw men.
But quite frequently, our goals are not to persuade each other. For example, if I speak publicly, it is not my goal of persuading you in particular, because I don’t even know who you are! I could also be aiming for more modest goals, like defending what I believe (without persuading others), or explaining why I believe what I believe, or exploring a topic to determine what I believe.
A lot of logical fallacies are based on a presumption of a structured argument. We agree on what we’re arguing about, we agree on what sides we’re taking, and we agree on what standards of evidence to apply. And that so often just isn’t the case.
Explanations of logical fallacies usually focus on what’s correct or incorrect. But to really practice critical thinking, we need to understand the social aspects of argument, including our respective goals.
i should keep a link to this in my back pocket. i’ve long had a similar view but couldn’t have formulated the expression of it as well as you have.
“There is obviously no way to “prove” that the subject matter is one thing or another.”
The subject matter of any argument is the disputed point, which is introduced by the person objecting to some claim.
(Can’t start an argument unless something is being disputed)
@John Morales,
What you’re saying, is that the subject matter is the subject matter, which is introduced by the person who introduced the subject matter.
🙂 Well, that’s one way to phrase it.
If someone makes a claim and another disputes it, then an argument exists; if nobody disputes it, then there is no argument.
In this specific case, I dispute that there is “no way to “prove” that the subject matter is one thing or another.”; the subject matter of our own argument is whether the subject matter of an argument can be “proven”. Not proven, but rather “proven”.
I say yes, there is.
A makes a claim; B disagrees. Either of them may at any time call the other a poopyhead, with or without supporting evidence.
At each step, the discourse bifurcates. A could have reiterated the claim, or could have criticized the disagreement; the accused poopyhead could defend or counterattack. We now have at least four issues on the table: the claim, the counterclaim, the respective characters of the disputants – eight or more positions.
Without a formal structure, such as strict courtroom or journal/blog protocols, very few debates will follow up on every fork of every path, as these will multiply literally exponentially. Even in such a framework, who could or would pursue every point to complete resolution? (It might make a good training exercise for an AI, with a very assiduous and patient controller, but everybody else would get bored &/or angry and go home.)
The process of picking and choosing which points to pound and which to let wither is sort-of tactical and Darwinian and sort-of a matter of personal predilection, both depending on the environment/audience of the debate. I would propose this takes us right back to competitive storytelling – but then, as many at FTB will confirm, I’m just a damned poopyhead.
@John Morales,
Sure, I’ll take your point. But let me explain what I meant.
If there is a disagreement in subject matter, i.e. I want to argue about X and you want to argue about Y, I cannot prove to you that you actually want to argue about X. Because it’s just false, I can’t prove a false proposition. Yes, I can point to how the argument was introduced, and “prove” the fact of the subject matter. But that doesn’t prove that you are interested in the subject matter.
Fair enough, Siggy. I do concur with the idea that most arguments are goal-oriented, not just academic.
There are also a lot of times when someone will present an argument about X, but they’re actually arguing Y. When you point out this dishonesty they’ll cry ad hominem.
A common one is when someone complains about racism and a white person says “that’s not racist”. The actual subject is whether the first person is credible in their testimony or if the white person should be considered more credible here.
if you point this out the white person has a sulk
It’s worth distinguishing between the formal and informal fallacies here (or more pithily, the ones that actually are hard rules and the ones that tend to get quoted on the internet). Denying the antecedent and affirming the consequent are reliably bad reasoning in a way that ad hominem, tu quoque, etc. just aren’t.
The part about one’s “goals” in an argument is important- superficially, it appears that the goal is to persuade the other person- but debate tactics like “fallacy spotting” don’t really work, if that’s what the goal really is.
I think it’s useful, at least in one’s own thinking, to think about what your *actual reasons* are for believing something. Like, people can make a lot of arguments about “here’s all the evidence to support my side” but often that’s not the actual reason they believe it. Sometimes when figuring out the actual reason for believing something, it turns out to be something embarrassing like “I don’t believe x because if x is true, that means the world is a scary place and I don’t want that to be true.” (I forget which fallacy this is- argument from consequences?) Or something that relates to your own experiences and isn’t going to be meaningful or convincing to anyone else. (And it’s good to be aware of fallacies because then you can recognize them in your own thinking.)
So obviously in an argument people wouldn’t actually say their actual reasons- you can only say your actual reasons in a setting where you can be honest and aren’t concerned about how people will view you. But the thing is, if you don’t get at the *actual reasons* no one will ever change their mind.
An example from creationism- if you get really into the weeds as a creationist, you learn a bunch of talking points about left-handed amino acids and right-handed amino acids. But nobody’s *actual reasons* for believing/ not believing in creationism is because of amino acids. So like, it’s pointless to have arguments about that.
This is why I try not to mock anti-vaxxers- like I mean obviously vaccines are the best thing ever invented and I place a high importance on my children not dying of measles- but I think it must be the case that people’s *actual reasons* for being anti-vaxxers are something about emotional needs that I can relate to. I don’t know, I’m not that familiar with their thinking, doesn’t really seem worth my time to get into that, but at least I don’t think I should make sweeping statements about how obviously ridiculous it is, because the *actual reasons* behind it all are probably relateable, even if they end up getting expressed in a way that’s really harmful.