Counterspells

Back in the early days of the internet, when New Atheism was king, people would collect and catalogue various logical fallacies. Then, when getting into an argument on some forum, they would identify and call out all the logical fallacies being used by their opponent; Ad Hominem! Appeal to Emotion! Tu Quoque! Post hoc, ergo propter hoc!

These days we tend to know better. Appealing to these lists of logical fallacies and their Latin names doesn’t help you, and it doesn’t help the conversation. The problem is that to the person you’re talking to, throwing around the names of these fallacies just sounds like an Argument from Authority—

Wait! I mean, it just sounds like you think you’re smarter or better educated than they are, which isn’t relevant to the argument. If you really had something to contribute, you would explain the problem with their reasoning, rather than just throwing out a Latin term they’ve never heard of. Even if they have heard of it, you should still take the time to tell them why you think it applies to what they said.

Logical fallacies are all sound complaints. But when disagreeing with someone, they deserve to hear the arguments those complaints are based on. More than that, you’re not going to convince someone of your position unless you actually try to convince them.

Fortunately, logical fallacies are such bad arguments that most of them can be rebutted in a few simple sentences. And the form of each fallacy tends to be so consistent that these rebuttals can be highly formulaic.


In accordance with the Rationalist tradition that requires everything to have a nerdy sci-fi or fantasy name, I am calling these formulaic rebuttals Counterspells.

Like I mentioned, I know that by now everyone knows how to avoid using these fallacies themselves. Most of us can also recognize when other people are using these bad arguments against us. Somehow the next step doesn’t come as easily.

You might think that an intelligent person, recognizing a bad argument they’ve seen a hundred times before, would be able to provide a knock-down counterargument. But for me at least, I still regularly crash and burn in my attempts to extemporize rebuttals. I see a lot of you all make the same problem.

Some writers have proposed some great Counterspells, but no one has collected them. To begin with, I’m going to review some excellent Counterspells by Paul Graham and Scott Alexander. I’ve lifted many of their examples directly, as these examples were carefully considered and strongly written; check out the links to see the original pieces.

Graham’s Counterspells

Paul Graham’s How to Disagree is an old favorite. Part of what I find so appealing about this essay is how Graham not only lays out several forms of unproductive disagreement, but includes clear examples of the sort of thing one should say in response.

Without further ado, Graham’s Counterspells:

Ad Hominem

Forms:

  • If there’s something wrong with [person’s] argument, what is it? If there isn’t, what difference does it make that [person] is a [member of category]?

  • The question is whether [person] is correct or not. If [person’s lack of authority or other issue] caused them to make mistakes, what are the mistakes? If there aren’t mistakes, [person’s lack of authority or other issue] isn’t a problem.

Examples:

If there’s something wrong with the senator’s argument, you should say what it is; and if there isn’t, what difference does it make that he’s a senator?

The question is whether I’m correct or not. If the fact that I’m just a college student caused me to make mistakes, what mistakes did I make? If there aren’t mistakes, then my age isn’t a problem.

Response to Tone

Forms:

  • It matters much more whether [person] is wrong or right than what their tone is.

  • I agree about [issue with tone], but I care much more about if [person] is incorrect somewhere. Could you point out where you think the argument is wrong?

Examples:

It’s clear that OP is pissed off about this issue, but it matters more whether OP is wrong or right than what their tone is. Do you think they’re wrong, or just an asshole?

Ok, maybe I was too flippant, but I still think my points are correct. If you think I’m actually wrong, could you point out where you disagree?

Contradiction

Forms:

  • I think [person] is just [stating the opposing case], without explaining why [person] thinks it’s correct. What evidence that convinced [person]?

  • It’s good to know [other person’s position], but I actually still disagree. But I need to understand what evidence makes you think [position].

Examples:

I think you’re just stating that you oppose gun control, without explaining why you think it should be opposed. What is it exactly about gun control that makes you so strongly against these policies?

Ok, it’s good to know that you’re concerned about genetically engineered crops specifically, but I actually still disagree. What evidence makes you think that these crops aren’t safe?

Alexander’s Counterspells

Noncentral Fallacy

One of my favorite pieces by Alexander is his description of his candidate for the Worst Argument In The World, what he calls the Noncentral Fallacy.

One thing that I like in particular about this piece is that he clearly explains how one should respond to this type of argument. This sets us up nicely for Counterspells.

Forms:

  • Yes, so what? [Object of discussion] is [member of a category]. But there are all sorts of other things we know about [object of discussion], and I think they’re also relevant to the question at hand. [Give example of other relevant information if possible.]

  • Normally when we think about [issue] we mean something like [typical/​central example of issue]. [This case] differs in that [ways in which this case is different].

  • Obviously [what we’re discussing] is an example of [issue]. The typical example of [issue] — [give typical example of issue] — is pretty [bad/​good/​etc.]. But a lot of the reasons [typical example of issue] — [list of 2-3 ways typical example of issue is bad/​good/​whatever] — don’t apply to the wildly atypical case of [what we’re discussing].

Examples:

The typical case of murder is Charles Manson breaking into a house and shooting someone. Abortion differs in that the victim is an embryo or fetus with less biological complexity and intelligence than the average rabbit.

The typical example of theft is someone mugging you in a dark alley and taking your pocketbook. Taxation technically qualifies as theft if you define the latter as “taking someone’s money through implied threat of force”, but it also differs from dark-alley-mugging in several important ways, like that it’s levied by a democratically elected government, that it’s supposed to be spent on useful programs, and that it’s collected in an orderly and predictable fashion. These differences seem to be important enough that most people support taxation even though they don’t support dark-alley-muggings.

Well, obviously. That’s kind of the point. And the typical example of racial discrimination—the Ku Klux Klan burning your house down or something—is pretty bad. But a lot of the reasons KKK-house-burning is bad—living in fear, locking downtrodden groups into a cycle of poverty, totally locking qualified people out of any job—don’t apply to the wildly atypical case of affirmative action.


Alexander has an essay called Varieties of Argumentative Experience, written as something of a response to Graham’s essay. This is excellent for my purposes, since it means that it has a very similar form. Here are some Counterspells drawn from this source.

Gotchas

Gotchas don’t share as much structure as other things on this list, so the form of their Counterspells are not as reliable as some others. I think that Alexander would also tell me that Gotchas aren’t worth engaging with, and maybe he’s right. But I hold out hope for people who are still playing gotcha, and a calm response can help to refocus the conversation on the issues, so here is an attempt.

You’ll notice that my approach here is to deadpan accept the fiction that they are making an honest contribution. Other approaches might also be possible.

Forms:

  • Well, [things] can vary along dimensions other than [thing being pointed out].

  • Yeah, but I still support [thing] because it has [these other consequences].

Examples:

Well, I don’t want to move to Cuba, but that’s because governments can vary along dimensions other than how big they are, and even though Cuba has a big government there are a lot of things about it that I don’t like.

Yeah, but I still support serious gun control measures because I think there are bigger issues than criminals having guns, like suicide rates and murders by private citizens.

Single Facts

Forms:

  • [Fact] admittedly does support your argument, but things with some bad features can be good overall. [Relevant example, if possible.]

  • [Fact] is a good point, but the issue is larger than that, and I think [fact] is outweighed by [facts that outweigh it].

  • Even if we all agree that [fact] is true, you’re only pointing out one [bad/​good] quality of [thing]. I don’t think it’s such a big deal. Why is [fact] really important?

Examples:

Trump lying about his grades admittedly does support your argument that he’s sometimes dishonest, but things with some bad features can be good overall. Trump can be dishonest as a businessman sometimes, but he has other good qualities.

Obviously foreign intervention does cost American lives, and that’s something I worry about too. But the issue is larger than that, and I think we should be prepared to make these sacrifices sometimes — things like genocide and religious dictatorships are much worse.

Even if we all agree that Hillary is crap at email security, you’re only pointing out one bad quality of hers. I don’t think it’s such a big deal. Why is her email security cred so important?

Single Studies

There are a lot of ways in which sharing a single study can be a nonproductive contribution to a discussion, so here are a few varieties of Counterspell.

Forms:

  • On any controversial issue, there are usually many peer-reviewed studies supporting each side. In just [however much time you spent searching] I was able to find [X number] articles with the exact opposite finding. If we want to answer this question, we need to take a look at the literature as a whole.

  • There are a lot of ways in which a study can be wrong, and it’s not always obvious when they are. Sometimes honest researchers make a mistake, or just get unlucky. Is there any other evidence out there for [thing] besides this study?

  • I think this study investigates [a much weaker subproblem; be as specific as possible] rather than [the larger problem].

Examples:

On any controversial issue, there are usually many peer-reviewed studies supporting each side. In just 15 minutes, I was able to find 7 articles (links here) that conclude that raising the minimum wage decreases unemployment. If we want to answer this question, we need to take a look at the literature as a whole.

I think this study investigates whether trigger warnings affect certain beliefs about trauma in the very short term rather than whether trigger warnings are helpful or harmful in college courses, which is the thing most people actually care about.

Isolated Demand for Rigor

This actually can be countered in a variety of interesting ways; see this essay for a more complete treatment. Most examples lifted from there.

Forms:

  • This is wrong because [clear counterexample to the demand for rigor]; [proposed rule] is a fake rule we never apply to anything else.

  • Presumably you think [thing that doesn’t pass the demand for rigor]. So why does [issue] have to justify itself according to rigorous criteria that nothing could possibly pass?

  • So you think that [rule stated explicitly]? Wouldn’t that imply [weird consequence they probably don’t support]?

  • You seem to be happy to [do thing that breaks the demanded rigor], but you switch to a position that [other thing isn’t ok because of demand for rigor]. Why the difference in opinion?

Examples:

Republicans have also been against leaders who presided over good economies and presumably thought this was a reasonable thing to do; “it’s impossible to honestly oppose someone even when there’s a good economy” is a fake rule we never apply to anything else.

Presumably there are lots of government programs you support (maybe PBS?), but you would never dream of demanding that we defund them in the hopes of donating the money to malaria prevention. But since you don’t support air strikes, suddenly that plan has to justify itself according to rigorous criteria that no government program that exists could possibly pass.

So you think that no one should ever be forced to pay for something they don’t like?Wouldn’t that imply liberals shouldn’t have to pay for wars? That seems like a weird consequence you probably don’t support.

You’re probably happy to talk about speed and blood pressure and comas and the crime rate, but somehow you switch to a position that we can’t talk about IQ at all unless we have a perfect factor-analytical proof of its obeying certain statistical rules. These seem equally rigorous, so why are you ok with one and not the other?

Disputing Definitions

Alexander correctly notes that this is one of the hardest issues to rebut; both because it’s easy to be seduced by this sort of argument, and because thinking critically about words and definitions is a very difficult skill to learn, a point not readily communicated in a few sentences.

Even so, here are my own attempts at a Counterspell for this issue:

Forms:

  • I’m not sure it matters whether or not [thing] is a member of [category]. People might even disagree with the definition of [category]. [Give example of definition drift if possible.] I think we both care a lot more about things like [actual issues with material consequences].

  • You can define [category] as [your definition] if you want, but that’s not usually what people mean by [category]. I don’t think factual or moral questions depend on how we use words.

  • You can define [category] as [your definition] if you want, but I’m much more concerned with [practical issue unrelated to the definition of a word].

(Note that this has a lot of overlap with the Noncentral Fallacy.)

Examples:

I’m not sure it matters whether or not being transgender is a mental illness. People might even disagree how to define what is and isn’t a mental illness. Some people still don’t think depression would count! I think we both care a lot more about things like whether or not transgender people should be allowed to change their legal gender, and whether they face discrimination.

You can define communism so that the USSR doesn’t count if you want, but that’s not usually what people mean by communism. I think the factual and moral issues stand, regardless of what words you use.

You can define war as “a formal state of armed conflict between state governments” if you want, but I’m much more concerned with the fact that people are dying in this conflict, whatever we call it.

Why can’t the chef call a tomato a vegetable? He just wants to think about whether or not to put it in a salad or on a burger. What’s wrong with that?

Other Counterspells

And here are a few of my own design, for some common fallacies:

Argument from Authority

Forms:

  • Just because [source] is [better educated/​more famous/​more expert/​smarter] than me, isn’t relevant to [the issue]. Sometimes experts are wrong and make mistakes. What arguments does [source] make that convince you of that position?

Examples:

Pinker being a Harvard professor and generally a smart guy isn’t relevant to his position on AI threat. Sometimes experts are wrong and make mistakes. What part of Pinker’s arguments do you think refute my position?

Straw Man

Forms:

  • Well I can’t speak for everyone else, but [straw man] isn’t what I believe about this case. Are you sure you understand what my position is? [If charitable, attempt to clarify your position.]

  • (If you are very confident) I don’t think anyone really believes [straw man]. Personally I believe [state your position]. Do you disagree with that?

Examples:

Well I can’t speak for everyone else, but I don’t actually want to ban all private ownership of guns. Are you sure you understand what my position is? All I’m suggesting is that we enforce the laws we have, and close some loopholes related to purchasing firearms at gun shows.

Tu Quoque

Forms:

  • It’s possible be against something and still sometimes participate in it. I may be a hypocrite about [whatever], but that doesn’t make me wrong.

Examples:

It’s possible be against something and still sometimes participate in it. Sure, it’s hypocritical for me to eat meat, but that doesn’t make me wrong about animal rights.

Sure, it’s hypocritical for me to smoke and tell you not to, but I’m still right that it would be bad for your health, and I can still want you not to start!

Appeal to Popularity

Forms:

  • The majority has been wrong all the time. [Pick one of the many examples, one that your audience will agree with.] I don’t care who [agrees/​disagrees with premise], I want to know whether [premise is correct/​incorrect].

Examples:

The majority has been wrong all the time. People used to think that the sun goes around the earth! I don’t care how many scientists believe in global warming, I want to know whether the evidence is strong enough to think it’s a real threat.

I don’t care if everyone on that subreddit thinks it’s right, I want to know why YOU think it’s right.

Come on, if all your friends jumped off a bridge, would you jump too?


There are, of course, many other things that can go wrong in the course of a conversation. There will be other concise, effective rebuttals, which can be added to this list.