Hide table of contents

People often ignore things which are long or otherwise would take significant effort to engage with. They’ll often engage anyway if they like it, it catches their interest, it triggers them, they find it appealing, or they have a bias in favor of engaging.

The overall result is a bias where long criticism is frequently ignored, especially if it’s reasonable instead of baiting or triggering. This means most people rarely engage with complex or sophisticated criticism. That doesn’t catch their interest well, and it’s too long or too much work to understand, so then they don’t put in the effort to engage. This systematic bias makes it harder for them to improve.

People often think that if some ideas were actually good or important, someone else would engage with it. When many people think this way, the result is systematic biases. The types of ideas which this bias negatively affects include ideas which are counter-intuitive, don’t give a good first impression, or seem like a lot of work to deal with.

People’s bias against long or difficult criticism frequently still applies to articles with a one paragraph summary near the start. The summary tends not to solve the problem. Sometimes people dismiss the arguments based on the simplified summary – they incorrectly try to use the summary to conclude that the criticism is wrong, even though they don’t really understand it. Sometimes they notice some nuance or detail isn’t covered in the summary, then assume it isn’t covered in the full article nor in the author’s mind (so if they asked a question about it, the author would already know the answer, but for some reason, good or bad, the author didn’t preemptively address that issue in the already-long article). Sometimes people use a summary to mistakenly assume a criticism is the same as something else they’ve heard before based on some partial similarity (or sometimes they correctly recognize that they already know it).

Sometimes people fear responding to the summary, without reading the rest, because they don’t want to say something which is addressed later in the article and therefore appear lazy or foolish. Whatever they have to say about the summary has a decent chance to be preemptively answered in the article somewhere, so saying it is risky. Skimming and search can help, but most people aren’t very good at those skills, and even if done well you can miss stuff.

If you respond to a summary, people may think you should have read the article and found the answer to your issue that was already there. You might get a reply referring you to the article section which answers you, at which point you would be expected to read that section and write another message, which is work you might not want to sign up for. Also, the section may not be independent of the other sections – to understand it well, you might have to read all sections before it or maybe the whole article.

So, people need methods to engage with longer criticism, so they don’t just biasedly ignore most of it. The methods need to respect their time, effort, chance of looking foolish, and other concerns. As far as I know, little practical guidance has been offered for this issue. Most people don’t seem especially interested in solving this problem. There doesn’t seem to be much demand for a change. One might suspect many people don’t really want to engage with criticism, which would explain why they aren’t seeking better methods to enable engaging with more criticism.

For those who would like to engage with criticism (or just with different or disagreeable ideas) more, I’ll share two methods you could use. Many other methods could be developed. There’s a lot of room for experimentation and progress; this stuff is under-explored. These are just two options. They’re nothing like a complete representation of what methods could be used.

Besides particular methods, it also helps to develop skills like getting more effective at skimming, speed reading, remembering what you read, organizing what you read (mentally or with notes, outlines or diagrams), and writing replies. The better you are at all those things, the lower cost there is to engage or partially engage. It’s also easier to engage if you know your own ideas and arguments well and are a good, experienced writer. A lot of the difficulty people have with engaging with rival ideas or criticism is simply the difficulty of trying to formulate their own current ideas in writing.

Method: Reply to the First Important Error

In short, if a criticism is correct, you’d want to engage with it. The scenario where you’d rather not engage with the criticism is if it’s incorrect. You could also do more detailed prioritization by considering how impactful the criticism is according to its author. You shouldn’t judge how impactful it seems to you, decide it doesn’t matter, and move on. That would be biased. You disagree with the author about whether it’s true and what its consequences are. You shouldn’t just assume you’re right about the consequences. If the author persuaded you that the criticism is true, it’s reasonably likely he could also persuade you to change your mind about its impact. If you’re wrong about the truth of the criticism, there’s a good chance you’re also wrong about its importance.

Note: It’s possible to have a discussion about the importance or impact of a criticism first, given the assumption that the criticism is true. Starting there, rather than by discussing the whether the criticism is correct, is uncommon. It requires skill at abstract thought and some ability to keep issues separate and track what the current discussion is (rather than getting distracted arguing with the assumed premise). And it requires using your opponent’s claim as a premise and assuming it’s true, which most people are uncomfortable with. But it’s reasonable and useful, for people who have the skill, to sometimes discuss importance before truth.

So, if you want to engage with correct criticisms, what can you do? Look for errors. If you read some, and see no errors, keep reading! Don’t assume it’s incorrect when everything you’ve read so far is correct in your own opinion! Assuming it’s incorrect because you don’t like the conclusion, even though you have no refutation for any of the arguments so far, is biased.

In general, as long as you’re reading and finding no errors, it’s a good article so far and it’s worth continuing to read. There are some caveats and ways that could go wrong. But as a general rule of thumb, if you’ve read some and found zero flaws, it’s worth reading a little more. And if you still find zero flaws, that applies again. The biggest exception to this is relevance. Even if I don’t see any flaws in an article about football, I still don’t care about football. (And I don’t make assertions about football for the article to contradict or criticize – except my assertion that football is an optional interest and it’s fine for me to not care. An article criticizing that specific claim, and arguing that interest in football is mandatory for moral or rational people, would be relevant to me. I’ve never heard of such an article, I’d be very skeptical, and I’d expect it to be easy to find flaws in it. But if there was such an article and I couldn’t find any flaws in the beginning, then I’d keep reading and actually find that surprising and interesting.)

If you do find an error, now you have a reason to stop reading. You now believe the criticism is incorrect. The error you stop for should be important. If it’s just a typo, or something minor that wouldn’t change the conclusion, then you shouldn’t stop, since it doesn’t give you any reason to judge the conclusion as false.

So one method you might consider is read until you see an error, then you’re done. However, you might have misunderstood it. Or the author might have a counter-argument to your criticism. It could be later in the article or in his head and ready to share (he can’t preemptively include all his arguments in his article – even a long book can’t fit everything). The risk of a misunderstanding, or of other information you don’t have addressing your criticism, is much higher than usual because you haven’t read the whole article and also haven’t conversed with the author at all.

So what do you do? You say what the first error is. That gives the author a chance to point you to a rebuttal later in the article or respond with a rebuttal. At that point, if it doesn’t change your mind, you should have noticed a second error. The reason you weren’t persuaded should be a specific error. (What if you intuitively disagree, but can’t explicitly point out an error? I have a section about that later.) If there is no error in the rebuttal that you can see, then it should change your mind. So you could say what the second error is.

When should you stop if you keep pointing out errors and the author (or anyone else who agrees with him or her) keeps making counter-arguments? There are a lot of policies you could use. A simple policy is stop talking after you point out three errors (you read their reply to the third error, and if it doesn’t convince you then you either say nothing or you say bye).

Why point out three errors and then give up? Because it gives you three chances to change your mind. It makes you look for specific errors instead of dismissing something for unspecific reasons (this helps combat your own potential bias). It gives the author a few chances to engage with you and persuade you. It’s a limited amount of effort that respects your time, but it’s more engagement than typical, so even if you miss out on some important ideas because you didn’t engage more, at least you gave it a better try than most people do.

If everyone adopted a “three error limit” policy, I think it would improve the world significantly. It’s flawed but good enough to make a difference. Why is it flawed? Well, you’re ending the discussion because of three claims which are in dispute. You might be wrong. Stopping after three errors is a simple, imperfect compromise between truth-seeking and limiting the effort you expend. There are other approaches which may be able to get more truth-seeking benefit per cost. However, many other methods don’t have a fixed limit on cost, so it used incorrectly you can get a bad outcome, but if used well with high skill then they can potentially do better. (Impasse chains are an example of more complicated method. And more broadly any method with robust failsafes, where other people can escalate and correct you if you’re screwing up, has advantages but also potential downsides due to the lack of a hard limit on effort. Having a public debate policy is an example of a failsafe.)

People commonly think things are wrong without identifying any specific, important error. Then they’re dismissive. That is biased. Responding to the first error (up to three times max) helps a lot because it allows some engagement with critics and people from other (sub)cultures and it gets you to think in terms of specific errors and put them into words. One of the main reasons people don’t do this is they aren’t very good at spotting errors or writing their thoughts in words. It’d take them a long time to write a one paragraph statement of an error, and they’d find the writing process hard, and the writing wouldn’t be very clear. If that’s you, and you care about rationality and truth seeking, then I suggest improving your writing and thinking skills until you no longer have that excuse for ignoring critical articles.

What Is an Error?

People often think there is an error because an article claims X and they already believe X is wrong for some reason. Claims that you disagree with aren’t errors. That’s not what identifying a mistake is.

Examples of errors include faulty reasoning, a bad argument, an unexplained claim that needed an explanation, some incorrect facts. But simply having a different conclusion than you do isn’t an error. It may be due to errors, but you need to figure out where they went wrong in reaching the conclusion.

If they don’t explain their reasoning for a conclusion you disagree with, there are two main possibilities. One, they aren’t arguing their case and that’s an error. Or, two, their article is focused on something else. You disagree with something that isn’t their main topic. It’s common that people use premises you disagree with and don’t argue for their premises in the current article. There’s nothing wrong with that. You can respond by asking them to link or cite arguments for the premise you disagree with that are written down in a different article or book. The reasoning should be written down somewhere.

What To Start Reading?

No one has time to start reading every article published. There are way too many people in the world writing too many articles. I’ve been talking about how to deal with an article you’re looking at. But which articles should you look at at all?

If everyone kept the same policies for what articles they pay any attention to – notice exist, read the first few paragraphs of, etc. – then I think the policy I’m suggesting (look for the first error) would help. Improvements in what people read in the first place are possible but not necessary.

Some people already do pay a little attention to some criticism. Then they often dismiss it as bad and wrong without clearly identifying a specific error. That’s a mistake.

To further save time, you could also limit your comments about the first error in an article to authors who seem potentially interested in debate. A lot of authors are just going to ignore you, and that’s often predictable in advance. It’s good to test that sometimes. You might be miscalibrated in your estimates about who won’t reply to you. You should occasionally check, especially in borderline cases but also now and then in cases you’re highly confident about. You don’t have to do all the checking personally though: if you see someone else share a criticism and get ignored, and you see someone else request a debate and get ignored, then you can use that data.

Replying to Any Error

A variant method is to skim and reply to the most important, or first, error that you find. Skimming involves skipping text and sometimes reading out of order.

The main point is to reply to one error that you find quickly. It doesn’t really matter if it comes first in the article, although it tends to be easier to reply to earlier errors. Later parts of articles have dependencies and back-references to earlier parts, so engaging with them is more complicated. Earlier parts have less leading up to them, so they stand on their own more.

If you read up to some point and think something is an error, there is a risk that if you read the rest of the article you would change your mind. However, if you skim around and think something is an error, there is a risk that if you read the rest of the article before the error, you’d change your mind. The risk of changing your mind due to unread earlier text is higher than the risk due to unread later text (for the same amount of text). So replying to the first error – and reading all the text before it – is lower risk than skipping parts or reading parts out of order.

Intuitive Disagreement

What if you intuitively disagree with something but don’t know how to point out a specific error in English sentences? Intuition can be used in rational, explicit debates. See my article Intuition and Rational Debate.

For more information on this subject, see my related articles:

Some people think only explicit ideas in language are rational. Some people think that if you can’t explain your disagreement, then you should be assumed wrong. Some people think if you can’t argue your case, then you’ve lose the debate and should change your mind. I am not one of those people. I think those people are wrong. If any of them are reading this, I challenge them to debate. I’m good at explicit debate, and I will stand up for other more intuitive thinkers. Dismissing intuitive ideas is a type of biased bullying. If you’ve ever been pressured by people who make clever arguments, while you struggled due to inferior rhetorical skills, you may enjoy my articles about intuition.

Method: Ask If Author Would Debate

One of problems with reading a long article, and having a bunch of criticism of it, is the author usually won’t engage with you. You put in all this work and you’re ready to engage, but then they don’t want to.

A simple solution is, before reading the article, you ask the author if they’d want a serious conversation, a debate, or something like that. Then you only read the article if they say “yes”. In this way, you’ll only read maybe one out of ten long articles, which is a big time and effort savings. And for the ones you do read, you’ll have a much higher chance of getting a good discussion. Also, the ones you read will be more likely than average to be high quality, because authors who are willing to engage in discussion about their articles tend to be above average (especially authors who agree to discuss in advance and directly say that they will, instead of wanting to keep their options open so that if your criticism is too challenging they can silently ignore it).

You should probably develop good skimming skills and give the article a brief skim before you ask the author about a conversation. You should at least check that the article’s topics are relevant to you. Sometimes titles are vague, or are even misleading, about what most of an article is about. When considering relevance, ask yourself “If this were true – if the author were right about all his claims – then would it be relevant, important or interesting to me? Would that affect anything I do in my life?” It’s pretty common for people to incorrectly treat something as irrelevant because they believe it’s false, but that’s a bias. Relevance and falsity are separate issues. To avoid that error, imagine and assume it’s true when considering relevance.


There’s a widespread problem where people dismiss ideas as wrong without pointing out any specific error or even identifying an error in their own mind. There’s also a widespread problem where large groups of people are systematically biased against engaging with long, complex criticism, even though some of the most important criticisms are long and complex (and counter-intuitive, surprising, weird, etc.). I’ve suggested some methods that can help. This area is under-explored, seemingly due to lack of interest in doing better, so I think other useful methods could be developed fairly rapidly if people actually wanted to.





More posts like this

Sorted by Click to highlight new comments since:

I have an unusual perspective on this. I skim nearly everything, and rarely see the benefit of completionism in reading. And I don't benefit from the author being nuanced or give different views a balanced take. I don't care about how epistemically virtuous the author is being! I just care about learning usefwl patterns from the stuff I'm reading, and for that, there's no value to me knowing whether the author is obviously biased in favour of a view or not. There are more things to say here, but I say some of them here.

Btw, I appreciate your contributions to the forum, at least for what I've seen of it. : )

Thanks for this article, it provided good advice, I even wrote down some notes. 

I still think that you should do a summary of the start of your articles, using bullet points and stuff like that. Some people won't go past the summary and still criticize, but they weren't going to read the full stuff in the first place. I  think many people find a summary helpful because it allows them to know if the article is worth checking and has interesting themes and insights.

Plus I find summaries helpful to take notes in a faster way.

It also allows to highlight some sentences that I find really worth reading like "People commonly think things are wrong without identifying any specific, important error" or "Why point out three errors and then give up? Because it gives you 3 chances to change your mind."

Curated and popular this week
Relevant opportunities