There are not infinitely many ways to argue badly. There are maybe six. Once you can sort a bad argument into the right family, you stop being surprised by it, and you start seeing the same moves everywhere: in comment threads, in meetings, in the news, and in your own first drafts.
That last one matters most. Spotting a flawed move in someone else's argument is easy and a little smug. Catching it in your own, before you hit send, is the skill that actually makes you a better thinker. This guide is organized so you can do both. Read it once to learn the shapes. Read it again asking which ones you reach for when you're losing.
The families overlap at the edges, and a single sentence can belong to more than one. That's fine. The point isn't precise taxonomy. The point is recognition: a fast, accurate sense of "I've seen this before, and here's why it doesn't hold."
Family 1: Attacking the person, not the argument
The simplest way to avoid answering a claim is to talk about the person who made it. The claim sits there untouched while the conversation moves to character, motive, or background.
Ad hominem is the base case: dismissing an argument by attacking the arguer. "You're not a doctor, so your opinion on this study is worthless." Whether the speaker is a doctor has no bearing on whether the study says what they claim it says.
Tu quoque is the "you too" version. Someone points out a problem with your behavior, and instead of answering, you point out that they do it too. Their hypocrisy might be real. It still doesn't make your behavior fine.
Genetic fallacy attacks the source rather than the substance. "That came from a tabloid, so it's false." Tabloids print true things sometimes; prestigious outlets print false things sometimes. Where a claim came from is a reason to check it, not a reason to dismiss it unread.
The tell for this whole family: the argument on the table is still standing, and nobody has touched it. The conversation just walked around it to reach the person.
Family 2: Misrepresenting the position
This family wins by arguing against a position nobody actually holds. The fake version is always easier to beat than the real one.
Straw man is the classic. You restate your opponent's argument as a weaker, dumber version, then knock that down. "You want a small cut to the military budget? So you want the country defenseless." The fix is to restate their claim in words they would accept before you respond to it. You lose some rhetorical force and gain everything else.
Motte and bailey is the reverse trick, run on your own argument. You advance a bold, controversial claim (the bailey). When challenged, you retreat to a modest, obviously true claim (the motte), then act as if defending the modest one rescued the bold one. The two claims wear the same words and mean different things.
No true Scotsman defends a sweeping generalization by redefining it on the fly. "No real fan would say that." When a counterexample appears, the category quietly shifts to exclude it, so the claim can never be wrong. A claim that can't be wrong isn't telling you anything.
The tell: the version being attacked or defended isn't the version actually in play. Always check that the argument under fire is the one someone really made.
Family 3: Dodging and deflecting
These moves don't misrepresent the argument. They change the subject and hope nobody notices the original point went unanswered.
Whataboutism answers a critique with a different critique. "You're worried about our spending? What about the last administration?" The other side's spending might be a fair topic. It doesn't resolve the point that was raised.
Red herring is the general form: any irrelevant point dragged in to pull attention off the real question. It often sounds relevant, which is what makes it work.
Moving the goalposts keeps the subject the same but changes the standard. You provide the evidence someone asked for, and suddenly that evidence isn't enough; the bar has quietly moved further away. No amount of proof satisfies a target that relocates every time you reach it.
Shifting the burden flips who has to prove what. "Prove it doesn't work." The person making a claim owes the support for it. Demanding that doubters disprove it gets the responsibility backward.
The tell: you ask one question and get an answer to a different one, or the conditions for "satisfied" keep sliding.
Family 4: Distorting the evidence
Here the logic might be fine and the topic might stay put. The trouble is in the data, selected or stretched to say more than it can.
Cherry picking presents the evidence that fits and omits the evidence that doesn't. One cold winter doesn't settle a debate about long-term climate trends, but it makes a tidy chart if you crop the timeline.
Anecdotal evidence substitutes a vivid story for a representative sample. "My grandfather smoked a pack a day and lived to 95." One memorable case feels more convincing than a study of thousands, and is worth far less.
Hasty generalization draws a broad conclusion from too few examples. Two bad experiences with a brand becomes "their products are all junk." Maybe. Two data points can't carry that weight.
Survivorship bias looks only at what made it through and ignores what didn't. "Three founders dropped out and got rich, so college is optional for success." You're not seeing the thousands who dropped out and didn't, because nobody writes articles about them.
The tell: the sample is doing work it isn't large enough or representative enough to do. Ask what got left out of the picture.
Family 5: Bad logical structure
This family fails at the level of form. Even if every fact is true, the conclusion doesn't follow from them.
False dichotomy presents two options as if they were the only two. "Either we ban it entirely or we accept total chaos." Most real choices have a middle, and the trick is hiding it.
False equivalence treats two things as comparable when they differ in degree or kind that matters. "A parking ticket and a felony are both breaking the law" is technically true and practically absurd.
Circular reasoning uses the conclusion as its own support. "It's the best because nothing is better." The sentence loops back on itself and never touches the ground.
Slippery slope claims one small step leads inevitably to a disastrous end, without showing the steps in between. It isn't always a fallacy: if you can demonstrate the mechanism that drags A to Z, it's a real argument. The fallacy is asserting the slide without showing it.
False cause and its common form post hoc confuse correlation with causation. Post hoc reasons that because B followed A, A caused B. Sales rose after the rebrand, so the rebrand worked, ignoring the season, the ad spend, and the competitor who just folded.
The tell: write the argument as "premises, therefore conclusion" and check whether the conclusion is actually forced. Often a true-sounding setup leads nowhere in particular.
Family 6: Overwhelming and exhausting
The final family doesn't try to be correct. It tries to make engaging too costly to bother with, and treats your exhaustion as a win.
Gish gallop buries you in a flood of claims, each one quick to assert and slow to refute. By the time you've debunked the first three, ten more are on the table, and an onlooker scores it as "lots of unanswered points."
Sealioning is relentless, polite demands for evidence and explanation, performed as good-faith curiosity. Each question seems reasonable. The volume and persistence are the attack. The goal is to keep you defending forever while never engaging with your answers.
The tell: you feel buried or worn down rather than answered. The volume of the argument has replaced the substance of it.
Turn the lens around
You can use these six families two ways. The defensive use is obvious: when an argument feels off but you can't say why, run it past the families and you'll usually find the shape. That alone makes threads less confusing and a great deal less persuasive in the wrong direction.
The harder and more valuable use points inward. The next time you're writing a reply and feel that little surge of confidence, pause and ask which family you're standing in. Am I answering their actual claim, or an easier one I swapped in? Am I responding to the point, or changing the subject? Does my evidence really cover this, or did I just grab the convenient half? People rarely reach for these moves on purpose. They reach for them when they're losing and want to feel like they're winning. That's exactly when noticing pays off.
A bad argument you catch in your own draft never costs you anything. A bad argument you send costs you the one thing this whole site is built to protect: being worth listening to.
The fastest way to get fluent is volume. Read enough examples and the shapes become automatic, the way a chess player stops calculating common patterns and just sees them. Every entry on this site has a counter-example section, because half the skill is knowing when a move that looks like a fallacy is actually a sound argument wearing a suspicious coat.
When you're ready to go deeper, the full list groups every entry by category, and using fallacy.is in a discussion covers what to do once you've spotted one in the wild. Learn the families first. The names come easy after that.