Ignoring “Small” Errors

Attitudes enabling ignoring “small” errors makes it significantly harder for critics to get attention and make progress. Even if they point out an error, and they are correct that it’s an error, and people agree with them … that often isn’t good enough. That makes the job of the critic very hard.

Making Friends and Influencing People

For a critic’s criticism to have an impact in the world, they may have to consider options like social climbing, marketing, and other activities separate from truth seeking. They may spend time developing different skills instead of focusing on truth seeking skills like research, logic and critical thinking. The influence-related skills may actually conflict with truth seeking rather than merely being a time-consuming distraction (that’s controversial and debatable).

Some skills are dual purpose. Writing clearly helps both with truth seeking and being influential. Other skills, like making a positive first impression, establishing rapport with people, or avoiding taboos, are not (directly) useful for truth seeking.

Something people tend to like a lot and respond well to is similarity. They like their subculture, in group or tribe. There are two basic ways to satisfy people’s preferences for similarity. You can be similar to them or you can fake it. Both have downsides. Being similar tends to involve making some of the same mistakes they do and having some of the same biases. Faking similarity compromises your integrity. The alternative strategy of criticizing people’s desire for similarity usually doesn’t get a positive response, unless done in a superficial way which focuses the criticism on other less tolerant, more tribalist communities.

Small, Unimportant Errors or Big, Controversial Errors

We can loosely categorize all errors into two groups, or view errors on one spectrum.

“Small” errors tend to be viewed as pedantic or unimportant. They’re also more clear-cut, indisputable, easy to establish, hard to debate, etc. So they’re easier to point out (this is especially important the more the audience is hostile or biased, or lacks shared premises or shared background knowledge with the critic). But they’re harder to get anyone to care about.

“Big” errors tend to be viewed as important and high impact, but complex and debatable. They’re easier to get anyone to care about if they agree it’s an error, but harder to get anyone to agree actually is an error. They tend to be hard to explain and hard to reach a conclusion about in debate.

Are there “medium” errors which seem some somewhat important and are somewhat controversial? Yes, sure. So we could add a third category or use a spectrum.

The error spectrum goes from “small” errors on one end to “big” on the other. We could number it from 0 to 100. As numbers (error sizes) increase, error importance increases but the ease of explaining the error decreases. This spectrum has two traits with an inverse relationship.

The result is there’s no great point on the spectrum for a critic to use. What you’d like is an easy to explain and very important error. Instead, you can have an important but hard to explain error. Or an easy to explain error that people will perceive as low importance. Or you can have something in the middle with a mix of both problems but less severe amounts of each problem.

This spectrum is only an approximation. It is possible, occasionally, to find an error that’s really simple yet really important. But this spectrum helps explain the typical experience of critics. Important things tend to be more complicated. Less complicated things tend to be perceived as requiring only small, uncomplicated fixes, and therefore as not very important unless they actually change the conclusion. If you left a negative sign out of some math, that is a “small” error, but it could change the answer from 50 to -50 and thus have a big impact on the conclusion. Getting the opposite conclusion often leads to double and triple checking, which makes this kind of error less common in ideas shared publicly.

Solution: Care About Small Errors

I propose that people care more about errors instead of dismissing them as “small” (or minor, unimportant, negligible, etc.). If a critic points out a few small errors, then you should find it plausible that maybe he’s right when he points out a big (but complex and hard to evaluated) error on the same topic or make a claim that you disagree with.

In general, critics ought to start by pointing out a few small errors to establish they can get things correct, then point out a bigger error that’s more important. Starting with a big, important error makes less sense. The small errors can be addressed more quickly and can establish a bit of a track record for the critic. Complex issues involve many small parts, so being able to get lots of small things correct is relevant. Pointing out several small errors both shows that the critic has relevant skills (making his bigger criticism plausible enough to spend some time considering) and also shows that the person or group being criticized may not have been careful, precise, accurate or skilled enough (making it more plausible they’d be wrong about something important).

What often happens instead is people are very resistant to small criticism. They get defensive and try to dismiss it as unimportant. If it’s correct but not very important, people should respond with “Great, thanks, I agree; you’re right. I don’t see how this makes a significant difference to the conclusion, but any correct criticism is always appreciated. Do you have more criticism to share?” And then every time the critic gets something right, you listen a bit more. If they do a dozen small criticisms in a row, maybe you get less interested and say “Look, you already got my attention since you seem good at making correct criticism. You can keep doing criticisms like this and it’s a positive contribution. But do you have some that matter more, even if they’re more complicated and harder to explain? I’d be willing to try to understand a big criticism from you. If you just have a lot of small errors you want to clean up, that’s fine. But if you have a big point, that’d potentially be even better.”

You Don’t Know How Important “Small” Errors Are

In general, you can’t reliably tell how important an error is until after you fix it. Once you know the solution, you can see how much the error mattered. But before you know the solution, that would require predicting the future growth of knowledge, which is unreliable.

You can guess whether an error is important and do significantly better than random guesses. But when you actually go to fix an error, sometimes it will turn out to be harder to fix than you expected. Sometimes it’s connected to some other issue, so you have to go change a second thing. When you’re changing the second thing, you notice it’s related to two other things which now have to be changed too. You can easily get a big chain reaction to deal with based on one little starting point. Unless an error is very clearly isolated from all the other ideas and parts, it’s hard to know there won’t be a chain reaction.

Error Clusters

Errors are evidence of more errors. Errors tend to come in clusters. Whatever caused one error is often capable of causing other errors too.

Many errors are hard to find or see. That some errors are easier to identify is really good news for us. It helps guide us for where to look for better hidden errors.

Small, visible errors are usually the tip of the iceberg. If they didn’t exist, how would we find the submerged part of the iceberg?

You aren’t done dealing with an error until you understand what process caused the error and what other errors it would cause. You also need to fix the process so it stops repeatedly creating errors.

Approximately all errors come from processes that can repeatedly create errors rather than being single, isolated or random incidents. Viewing small errors in this context – that they’re evidence of a process that creates multiple errors and keeps creating errors over time – will help change your perspective on them.

If you care about error-creating mental processes, it’s important to not only value whatever evidence about them you can find (including “small” errors), but also to fix errors the right way. If you put a superficial (“bandaid”) fix on a “small” error, without fixing the cause, then you’re basically getting rid of visible evidence and making the root cause harder to identify and fix. In general, if you just change something to the right answer without knowing how or why the wrong answer was reached, that’s a superficial fix that doesn’t address the process that led to the error.

If It’s So Small, Just Fix It

If an error is really small, just fix it. Why resist fixing it? If it’s an easy fix, it’ll be easier to fix it than to argue with the critic who is correct. Are you just hostile to criticism? Or defensive and don’t want to let the critic have any small wins even when he’s right? Those are bad attitudes!

One reason people don’t want to fix errors is they lack any good place to keep records of errata, modifications to texts, etc. This isn’t that hard of a problem. You can start a topic on EA, like your short form, but for noting small corrections to ideas and literature that you endorse. For some of your own work, you can also edit it to fix errors, although you should generally keep records of what was changed, why, what the original said that was wrong, etc. You don’t want to hide that an error ever happened or delete evidence about it. For bigger more complex corrections, you can write a separate post about them and put a link to it in your error correction topic. You could also just have a single post that you edit with updates that links to all the corrections in various places like posts you wrote or nested comment threads where you discussed with critics.

If anyone else cares about any of the same errors that you do, and they do a correction that satisfies you, you can link to theirs instead of writing your own. If lots of people do this kind of thing, they can all save each other a lot of time. Also, instead of linking to each error correction individually, you might link to a collection of error corrections and endorse the whole group. Someone might organize 50 error corrections on one topic in a post, and you might be satisfied with all of them and link it. Or if you’re satisfied with all but two, you can link it and note the two exceptions.

Most people aren’t trying to keep track of correct ideas in an organized way. Maybe that’s fine but at least some people trying to be thought leaders or public intellectuals should do this kind of thing.

Caveat: Typos and Grammar

Typos which don’t cause relevant ambiguity and aren’t within quotes aren’t a big deal, as long as they aren’t very frequent. Infrequent grammar errors which don’t cause ambiguity also aren’t a big deal. People often seem to assume that many other “small” errors, with different causes, should be treated like typos – like random, arbitrary things that just happen sometimes.

The reason typos and grammar errors usually don’t matter is we know their source or cause. The post mortem is simple. In short, they aren’t useful evidence to find other errors of other types. They just mean the author could have spent more time practicing typing or grammar, or could have done more rereading during and after editing (a lot of typos aren’t from initial typing, but are instead due to editing something while failing to edit something else connected to it, thereby creating an inconsistency).

Most small errors aren’t like typos and grammar mistakes. Many wording errors that most people would see as “small” are actually related to topical ,conceptual errors by the author. Those errors merit investigation to understand what the incorrect concepts are and what other errors those incorrect concepts led to.

Factual or citation errors are also important because they indicate the author is sloppy, and probably made more errors of the same type. And these kinds of errors enable bias. People who get facts wrong tend to get them wrong in ways that favor their conclusion, not in ways that harm their conclusion.

Read more in my article “Small” Errors, Frauds and Violences.