Ask HN: What logical fallacies irritate you the most on Hacker News? | Hacker News Reader

Ask HN: What logical fallacies irritate you the most on Hacker News?

18 pointsamichail26 comments

One I find annoying is when someone gives a counterexample as if that would be enough to shoot down an (implicitly) statistical claim/argument.

View on HN

I'm not a fan of people trying to shoehorn arguments (especially: arguments they are losing) into strict logical exercises. It's like a "work to rules" strike, message board style. The point of a message board argument is to influence and communicate and, sometimes, to arrive at a shared understanding of the truth (some topics, subjective ones, don't readily admit to a convergence on "truth" and yet make perfectly fine fodder for an HN discussion). They don't all need to be scored from the Nizkor list.

I'm not saying logical fallacies are all benign, but if you ask me what fallacy annoys me most on HN, it's the fallacy that logical fallicies are themselves invariably germane. They aren't.

I wonder if there is a conflict here because for some people careless statements are reliably at too low a precision to be helpful because one has already considered the topic at higher precision.

When you are an expert at something, to improve further you need to hunt down every last mistake you can find. You don't just dismiss logical fallacies as not being germane. Sometimes they don't look germane but actually are for a subtle reason.

By contrast, for newbies, even if a fallacy is germane in some subtle way it doesn't matter b/c they aren't learning subtleties.

I would like to hold my hands up and apply for penitence here, I am definitely guilty of that. A lot of it has to do with the fact that this is one of the few places on the internet where the majority of people are capable and qualified to have these "work to rule" (lovely analogy) type of logical debates. In this environment I have a tendency to debate in this way simply because I find it very intellectually challenging and rewarding. It also may well come from the fact that I and others often comment here as a break from coding, and as such we are in a very rigid, strict logical mindset when we post.

(I would hope however that I never use it as a get-out-clause in an argument I am losing. That is simply intellectually dishonest.)

Also it is perhaps because I am still reading Godel Escher Bach, which is shot through with this kind of thinking, and still enjoying it immensely. Large tracts of it are in fact made up of the very type of debates you describe. This sounds like awful reading, but bear in mind they are written in the style of Lewis Carrol.

https://secure.wikimedia.org/wikipedia/en/wiki/G%C3%B6del,_E...

I'm not sure if this is a logical fallacy per se, but I have observed some very strong groupthink effects with respect to voting dog piling, both for positive and negative votes. Strongly postive comments and stories tend to have accelerating up voting rather than the expected fixed frequency of votes, implying that number of votes is itself a factor for some in deciding whether to vote for something. It shouldn't be.

Yeah, it's increasingly easy to find out which side of an argument has the more popular viewpoint, even if both sides are making good points and intelligent comments.

None. I don't care what people say or how they say it because we're all different.

However, since it just came to mind, on StackOverflow.com, I despise when someone pops in and says "read the documentation" then provides a link. As if I hadn't already read it a thousand times and ended up at Stack because I was stuck.

Whoa. Way off topic.

One logical fallacy I don't like is the Sunk Cost fallacy. It says that because I already read half a book, I should read the rest too, even though I don't want to, b/c that will somehow make reading the first have less of a waste. It's like throwing good money after bad.

Another example is when you write a draft comment, that is a sunk cost. When you recognize that comment is off topic and shouldn't be posted, you cannot recoup your sunk cost by posting it anyway. That just makes things worse.

You'll have 150% better luck avoiding those comments if you phrase your question in terms of your understanding (or not, as the case may be) of the documentation that you (apparently obviously) already read. People don't know your unmentioned history.

This may be a little off topic, but downvoting w/o a comment seems rude. Let me help you balance that a little.

In response to your SO complaint: I will sometimes reply with a request to read the documentation (including a link to the relevant section), but only when it's very clear that the OP has not read the docs. It's especially prevalent for me, b/c I tend to stick to the PHP/lowest common denominator languages.

> downvoting w/o a comment seems rude

Happens ALL the time. I generally prefer to leave comments rather than downvote too. But most people don't like to write anything. If explanatory comments were supposed to be used, there wouldn't even be downvotes at all.

PS I did not downvote him even though I left a negative comment.

My policy is to down vote or comment, but not both. If it's worth a reply, why would I down vote it? If it doesn't contribute to the discussion, why would I reply?

Because by replying, you can point out flaws or disagreements in the parent post. That helps the person who wrote it get a little better.

Additionally, downvoting w/o a comment feel punitive, as opposed to corrective. It also seems a little lazy -- "I'll click the downvote arrow, but won't take any time to help them improve."

If I believe there's a good reason to help them improve, then I can do so and there's no need to also down vote, as my feelings will be on display for everyone to read.

Also, most posts with flaws do contribute to the discussion, I just disagree with the reasoning. That's not the right place for a down vote. IMO.

A down vote might be used for a situation where someone else has already explained why the comment is not helpful, or perhaps something beyond redemption like "Pwned, LOL" :-)

>If it's worth a reply, why would I down vote it? If it doesn't contribute to the discussion, why would I reply?

This is very insightful, and I will be keeping this notion at the front of my mind when posting. Although I think this mirrors my natural behaviour to an extent anyway.

Thinking about it, this is probably part of the HN karma/ranking "secret sauce" anyway. If not, it should be.

It depends. Sometimes I see comments that I don't understand are downvoted - I upvote these and sometimes comment.

But most downvoted comments are downvoted for an obvious reason: they're snarky, personal, ad-hominen attacks, trolls or just plain wrong. It wouldn't add anything to the discussion to comment. Here's an example of a snarky comment which deservedly got downvoted for obvius reasons: http://news.ycombinator.com/item?id=2038245 (it's mine - and I apologise...)

Totally disagree. I regularly down-vote ad hominem, jack assery, and anybody complaining about down-voting. I will never comment because I refuse to lower myself to their level.

You have to understand, there's a vast flood of people at Stack Overflow who haven't read the documentation, or at least not with any care. Every time I see a post asking, "Why is my code leaking when I write `self.foo = [[Thing alloc] init]`?" I know that person has not taken the time to read the Cocoa memory management guide. Reading that guide will benefit those people a lot more than me just answering their one question while they remain ignorant of the topic as a whole.

So if someone links you to some docs, their brain probably just incorrectly pattern-matched you as that very common kind of poster.

"Whoa. Way off topic."

Maybe not. It's not uncommon here for people, when they encountering a post they disagree with, to offer their interpretation, tossing in "Read the [article|post|blog|whatever]" as their, um, argument.

Sometimes (maybe mostly) this is done as a deliberately smug insult. But other times it seems the poster really believes that there is but one plausible interpretation of something, and that the only way someone could not have the same interpretation would be to have not read to item in question.

Either way it fails to improve the discussion because typically it dismissively ignores any accompanying argument.

Windbaggery, a.k.a. the fallacy of thinking that wrapping your ideas in intellectual-sounding verbiage will make them invulnerable.

It seems like philosophy students see this far more than other people (based on my informal survey of my philosopher friends) but confusion of descriptive and normative statements, also known as David Hume's "is-ought" problem, is one of the most common fallacies I see anywhere.

This is most easily seen on the more populist articles on Hacker News (which I simply found by going to /best).

Here's an example: http://news.ycombinator.com/item?id=2037139

These people are having an argument about whether it is or is not slavery i.e., a descriptive judgment. However, the question of whether or not it is slavery has little bearing on whether or not it should be done. The commenters are confusing a descriptive statement as relevant to a normative question.

Although what you say does happen, I don't see how that comment thread an example of it. Somebody claimed it was slavery, and other people argued that it was not. To say that particular line of conversation was about whether it should be done is a better example of the fallacy — you link to a descriptive debate, and because you think it should be a normative debate, you say that they were arguing the wrong thing.

It's true that I'm counting relevance to the topic in the conversation. Given that these are replies to a thread with a normative statement, the responses are either a fallacy or completely off-topic. I chose the more charitable interpretation.

As far as I can tell, it was neither fallacious nor completely off-topic. It was a response to the logical argument, "We should not do this because this is slavery, and we know slavery is wrong." This is a valid argument, so they were replying that the argument is factually unsound, which is a descriptive sub-issue within the larger normative question. If we are not allowed to debate the facts and can only debate the normative conclusions, arguments almost invariably devolve into emotional shouting matches.

Actually, I think that whether or not something can be called "slavery" is irrelevant to whether or not it is wrong. If it shares the same moral violations that our conventional idea of slavery does then it is wrong, but not because it is called slavery but because it violates the same moral or ethical precepts. If it could be called slavery but doesn't violate these precepts then it is not a moral wrong. This is actually slightly related to Socrates' question in Plato's Euthyphro.

Wishful thinking. For instance, a recurring theme is to predict a given platform will die (Flash, desktop apps, Apple's App Store, relational databases, etc.). Although I often find myself agreeing with those predictions, the arguments presented often take the form of a wish rather than an evidence based observation.

Appeal to auth- Wikipedia.

I hate the use of buzz words and namedropping "theories", and this is a close relative.

Just because there's a friggin' article about something vaguely tangential to the discussion does not mean that you've refuted someone's argument by linking to it. Five bucks says you haven't even read the article, but just used it based on the title.

Sometimes people don't even understand the science behind it, and sometimes they've misconstrued the actual concept of the theory.

Some people just toss in a link like a large, dead trout on the kitchen table as their entire comment without explaining their - supposed - train of thought. But hey, you want people to know that you've learnt a fancy word.

Reply on news.ycombinator.com