Hannah suggests at the family dinner that maybe everyone should eat a little less meat. Her brother Dmitri answers in about forty seconds:
"Says the person who flew to Lisbon last month. Look, either we all go vegan or we stop pretending, right? Grandpa ate steak every day and lived to ninety-one. Everyone I know eats meat. And if we start cutting meat, next they will tell us what car to drive. You can't prove it's actually bad for you, anyway."
Nobody at the table can quite say what went wrong, but something clearly did. Dmitri just packed six logical fallacies into one short speech: tu quoque, false dilemma, hasty generalization, bandwagon, slippery slope and an appeal to ignorance. Hannah might be wrong about meat. But nothing Dmitri said actually shows that.
This guide is a working list of 25 common logical fallacies, grouped by how they break an argument, each with a realistic example and a short answer you can use. Then comes a cheat sheet, advice on calling fallacies out without sounding like a hall monitor, and a way to train the skill with AI debaters.
What a logical fallacy is (and is not)
A logical fallacy is a flaw in reasoning: the reasons offered do not actually support the conclusion, even though they sound like they do. A fallacy is not a false fact. It is a broken bridge between the facts and the claim. Aristotle cataloged a set of them in Sophistical Refutations, and people have been adding to the list ever since.
Formal vs informal fallacies
A formal fallacy is a problem with the structure itself: "All cats are mammals. My dog is a mammal. So my dog is a cat." It fails no matter what words you plug in. An informal fallacy is a problem with the content or assumptions: the argument leans on something irrelevant, ambiguous or unproven. Almost every fallacy you meet in real life is informal, so that is what this list covers.
A bad argument can have a true conclusion
"Grandpa ate steak and lived to ninety-one, so meat is fine in moderation" is broken reasoning, yet the conclusion might happen to be true. A fallacy tells you the argument failed. It does not tell you the opposite is correct.
Assuming a claim is false because someone argued for it badly is itself a fallacy, called the fallacy fallacy. "That's a slippery slope, so you're wrong" does not follow. The right conclusion is smaller: "That reason doesn't work. Do you have a better one?" Spotting a fallacy is the start of a better conversation, not the end of it.
Fallacies of relevance
These arguments bring in something that feels connected to the question but does not actually bear on whether the claim is true.
1. Ad hominem
Attacking the person instead of their argument.
Example: In a planning meeting, Grace argues the release date is too aggressive. A colleague says, "You've only been here six months."
How to answer: "Maybe, but which part of the timeline do you think I have wrong?"
2. Tu quoque ("you too")
Dismissing a criticism because the person making it is inconsistent or hypocritical.
Example: Dmitri's "Says the person who flew to Lisbon."
How to answer: "Fair, I'm not perfect on this. But the question is whether the point is true, and that doesn't depend on me."
3. Appeal to authority
Treating a claim as true because a respected person said it, especially outside their field.
Example: "A famous tech founder said college is a waste of time, so it is."
How to answer: "What was their reasoning? And is this their area?" Citing genuine expert consensus is fine. Borrowing prestige is not.
4. Appeal to emotion
Using fear, pity, anger or guilt in place of reasons.
Example: A news segment about a new traffic rule opens with a tearful interview and never explains whether the rule would have prevented the accident.
How to answer: Acknowledge the feeling, then ask for the link: "That's awful. What makes us think this policy would change it?"
5. Bandwagon (appeal to popularity)
Claiming something is true or good because many people believe or do it.
Example: "Everyone on the team uses this framework, so it's obviously the right choice."
How to answer: "What does it do better than the alternatives for this project?"
6. Red herring
Changing the subject to something that distracts from the original question.
Example: Grace asks her landlord why the heating has been broken for two weeks. He replies with a long complaint about how expensive repairs have become.
How to answer: "I understand costs are up. When will the heating be fixed?"
7. Genetic fallacy
Judging a claim by where it came from rather than on its merits.
Example: "That idea came from the marketing team, so it's probably fluff."
How to answer: "The source is worth checking, but let's look at the idea itself. Where does it go wrong?"
Fallacies of presumption
These arguments sneak in an assumption that has not been earned, or reshape the debate so one side cannot lose.
8. Straw man
Replacing someone's actual position with a weaker version that is easier to knock down.
Example: Hannah says, "Maybe eat a little less meat." Dmitri replies as if she said, "Ban all meat."
How to answer: "That's not my position. I said a little less, not none." The best long-term cure is the opposite habit, steelmanning.
9. False dilemma
Presenting two options as the only ones when others exist.
Example: "Either we hire a senior engineer now or we miss the deadline." What about cutting scope or moving the date?
How to answer: "Are those really the only two options?" Then name a third.
10. Begging the question
Using the conclusion as one of the premises, so the argument goes in a circle.
Example: "This source is trustworthy because it says it only publishes the truth." (Note: people often say "begs the question" to mean "raises the question." The fallacy is the circular one.)
How to answer: "That assumes the thing we're trying to establish. What's the evidence from outside the claim itself?"
11. Loaded question
Asking a question that contains an unproven assumption, so any direct answer accepts it.
Example: In a performance review: "Why do you keep missing deadlines?" If Grace missed one deadline in a year, "keep" is doing a lot of work.
How to answer: Answer the assumption, not the question: "I'd push back on 'keep.' I missed one, in March."
12. No true Scotsman
Protecting a generalization from a counterexample by redefining the group on the spot. The name comes from philosopher Antony Flew's example.
Example: "No real runner uses a treadmill." "My marathon coach does." "Well, no real runner."
How to answer: "Let's define 'real runner' first, because the definition keeps changing."
13. Special pleading
Applying a rule to everyone else but making an exception for yourself or your side, without a good reason.
Example: A manager insists everyone must be in the office three days a week, but works from home on Fridays "because my situation is different."
How to answer: "What makes this case different in a way that matters?"
14. Moving the goalposts
Raising the standard of evidence every time the previous one is met.
Example: Dmitri says he'll believe remote work is productive when he sees data. Hannah shows him her team's sprint numbers. Now he wants a peer-reviewed study.
How to answer: "What would change your mind? Let's pin it down now."
Causal and statistical fallacies
These are the ones that thrive in headlines and social media, because they look like data.
15. Post hoc ergo propter hoc
Assuming that because B happened after A, A caused B.
Example: "I started taking vitamin C and my cold went away in a week." Colds usually go away in about a week anyway.
How to answer: "What would have happened without it? Is there another reason B followed A?"
16. Correlation vs causation
Treating two things that move together as if one causes the other.
Example: A news piece notes that people who drink more coffee report more productivity at work. Or maybe busy, driven people simply drink more coffee.
How to answer: "Could something else cause both? Could it run the other way?"
17. Slippery slope
Claiming that one step will inevitably lead to an extreme outcome, without showing why each step follows.
Example: "If we let the kids have phones at dinner once, they'll never talk to us again." Some slopes are real, but each link needs its own reason.
How to answer: "Which step in that chain is most likely, and what would stop it?"
18. Hasty generalization
Drawing a broad conclusion from a small or unrepresentative sample.
Example: Grandpa ate steak daily and lived to ninety-one, so meat is harmless.
How to answer: "That's a real example, but it's one. What does it look like across many cases?"
19. Cherry picking
Selecting only the evidence that supports your case and ignoring the rest.
Example: A startup pitch deck shows its three best months of growth and leaves out the flat year around them.
How to answer: "Can we see the whole dataset?"
20. Gambler's fallacy
Believing that past random events change the odds of future independent ones.
Example: "We've had four girls, so the next one's bound to be a boy." Or in sales: "I've been rejected nine times today, so the next call is due to be a yes."
How to answer: "If each event is independent, the odds don't remember what came before."
21. Texas sharpshooter
Finding a pattern after the fact by focusing on the hits and ignoring the misses. The name comes from the joke about a man who shoots at a barn, then paints the target around the tightest cluster of holes.
Example: A pundit claims a set of predictions "came true," listing the five that did and forgetting the fifteen that didn't.
How to answer: "Was the prediction made before the results? How many didn't hit?"
Fallacies of language and burden
These twist words or shift responsibility for proving a claim.
22. Equivocation
Using a word in two different senses within one argument, as if they were the same.
Example: "The law says we have the right to free speech, so it's right for me to say whatever I want at work."
How to answer: "I think we're using that word two ways. Which meaning do you intend here?" Defining terms early (see how to debate) prevents most of these.
23. Appeal to ignorance
Claiming something is true because it has not been proven false, or false because it has not been proven true.
Example: Dmitri's "You can't prove it's bad for you." A lack of proof either way is not evidence for his side.
How to answer: "Not being disproven isn't the same as being shown. What positive evidence do we have?"
24. Burden of proof shift
Making a claim and then demanding the other person disprove it.
Example: Online: "This supplement boosts memory. Prove it doesn't."
How to answer: "You're the one making the claim, so I'd need to see your evidence first."
25. Middle ground (argument to moderation)
Assuming the truth must lie halfway between two positions.
Example: One colleague says the report is due Friday, another says the client asked for Monday. "Let's split the difference and say Saturday."
How to answer: "Compromise works for preferences. For facts, let's find out which one is actually right."
Logical fallacies cheat sheet
Here is the whole list compressed, for when you need it fast.
| Fallacy | What it sounds like | How to respond |
|---|---|---|
| Ad hominem | "You're too junior to have an opinion." | "Which part of my point is wrong?" |
| Tu quoque | "You do it too." | "Maybe, but is the point true?" |
| Appeal to authority | "A famous person said so." | "What was their reasoning? Is it their field?" |
| Appeal to emotion | "Think of how awful it would be." | "Agreed it's awful. Would this fix it?" |
| Bandwagon | "Everyone's doing it." | "What makes it right for us?" |
| Red herring | "Speaking of which..." | "Let's come back to the original question." |
| Genetic fallacy | "Consider the source." | "Let's judge the idea itself." |
| Straw man | "So you're saying..." (something you didn't say) | "My actual position is..." |
| False dilemma | "It's either X or Y." | "What about option Z?" |
| Begging the question | "It's true because it's true." | "What's the evidence from outside the claim?" |
| Loaded question | "Why do you always...?" | "I'd question the 'always' first." |
| No true Scotsman | "No real X would do that." | "Let's define X before we continue." |
| Special pleading | "The rule doesn't apply to me." | "What makes this case different?" |
| Moving the goalposts | "That's not enough evidence." (again) | "What would change your mind?" |
| Post hoc | "It happened after, so it caused it." | "What else could explain it?" |
| Correlation vs causation | "They go together, so one causes the other." | "Could a third factor cause both?" |
| Slippery slope | "This leads straight to disaster." | "Which step is actually likely?" |
| Hasty generalization | "I know someone who..." | "What about across many cases?" |
| Cherry picking | "Look at these results." | "Can we see all the results?" |
| Gambler's fallacy | "It's due to happen." | "The odds don't remember." |
| Texas sharpshooter | "Look at the pattern!" | "Was it predicted before?" |
| Equivocation | Same word, two meanings | "Which meaning do you mean?" |
| Appeal to ignorance | "You can't prove it's not." | "What positive evidence is there?" |
| Burden of proof shift | "Prove me wrong." | "Your claim, your evidence." |
| Middle ground | "The truth is somewhere in between." | "For facts, let's check." |
Some of these are genuinely contested in practice. Is citing a doctor about your health an appeal to authority? Try this one with two sides:
How to call out a fallacy without sounding pedantic
Knowing the names is the easy part. Using them in a real conversation without making people defensive is harder. Saying "that's a tu quoque" at a family dinner rarely wins anyone over.
Name the damage, not the Latin
Instead of labeling the fallacy, explain what it does to the argument. Compare:
- "That's a straw man." (Accurate, but it sounds like a penalty flag.)
- "I think that's a stronger version than what I said. I only suggested eating a little less meat." (Same point, and it keeps the conversation going.)
Ask a question instead
Most fallacies can be exposed with one honest question. "Are those the only two options?" handles a false dilemma. "What would change your mind?" handles moving goalposts. "Is there another reason that could explain it?" handles post hoc. Questions let the other person find the gap themselves, which is far more persuasive than being told. If you want to get better at this, Socratic questioning is essentially a toolkit for it.
Grant what is true first
Many fallacious arguments contain a real point. Hannah could say: "You're right that I fly more than I should. And I don't want anyone telling you what car to drive. My point is just smaller than that." Conceding the fair parts makes your pushback far more credible.
Also, pick your battles. Casual conversation is full of loose reasoning. Save your energy for when a flawed argument is driving a real decision.
Catch them in your own reasoning
The most useful fallacy detector is the one pointed at yourself. A few habits that help:
- Write your argument in one line: claim, because, reason. If the "because" is about the person, the crowd or a single anecdote, you have a problem.
- Ask what would change your mind. If the honest answer is "nothing," you are probably special pleading or moving goalposts without noticing.
- Look for the misses. When you notice a pattern that supports you, deliberately search for cases that don't fit.
- Argue the other side. Build the best case against your view before you defend it. A devil's advocate exercise catches your own cherry picking fast.
- Watch for strong feelings. When you really want something to be true, slow down.
Here's a good one for testing your own reasoning, because most people have a strong gut reaction to it:
Train fallacy spotting with AI debaters
Reading a list of fallacies is like reading about swimming. You only learn by getting in the water, ideally with arguments you are not emotionally invested in. AI debates give you that. On Debatly, several AI models and characters argue different sides of a question across rounds, and a judge weighs the arguments and gives a verdict with reasoning. That gives you a steady supply of arguments to dissect, on any topic you choose. If the idea is new to you, what an AI debate room is walks through how a session runs.
A simple practice routine
- Pick a charged topic. Something where people tend to reach for shortcuts: diet, remote work, school policy, a news story. Run it in the debate room.
- Read each round with the cheat sheet next to you. Mark every place where you think a speaker leaned on popularity, an anecdote, a false choice or a shifted burden. Write down the fallacy and the sentence.
- Compare with the judge. Read the verdict from Judge Alex and see whether the reasoning the judge rewarded matches what you thought was strong. Where you disagree, work out why.
- Bring in a skeptic. Take the argument you found most persuasive and put it in front of Viktor, the skeptic, whose job is to challenge evidence. Ask him directly to flag any weak reasoning or fallacies. Then check his list against yours.
- Rewrite one argument. Take the weakest point you found and rewrite it without the fallacy. If it falls apart once the fallacy is removed, that tells you something about the original.
Grace used a version of this before a budget meeting. She ran "Should our team move from annual to quarterly planning?" as a debate, then asked Viktor to pick apart the side she agreed with. He pointed out that her favorite argument ("three companies I admire do this") was a bandwagon, and that her proof of success was a single good quarter. She went into the meeting with the same position and two better reasons.
For variety, a Devil's Advocate debate puts Petra on the attack against your view, and Socratic mode (or the one-on-one Socratic method AI) swaps the opponent for a questioner, ideal for catching circular reasoning in yourself.
An honest limit: AI debaters and judges can miss fallacies, over-label them, or make them. Treat their flags as a second opinion to check, not a final ruling. That checking is the skill you are building anyway.
Watching two AI models argue is also a good reminder that fluent, confident writing can still hide a broken warrant. Try it on a topic that pulls slippery slopes out of both sides:
Start spotting them this week
You do not need to memorize all 25. Start with the six that show up most in everyday arguments: ad hominem, straw man, false dilemma, slippery slope, hasty generalization and post hoc. Listen for them in one meeting and one news story this week, and try to answer one of them with a question rather than a label.
Then practice on arguments where nothing is at stake. Open a debate on Debatly, read it with the cheat sheet, and ask Viktor to check your work. Try for free, and pick a topic you already have a strong opinion about. That is where the fallacies hide best.
“Logical Fallacies: 25 Common Ones, With Everyday Examples”
