Logical Fallacies
Recognizing flawed reasoning patterns and knowing how to respond when you encounter them.
What Is a Fallacy?
A fallacy is a flaw in reasoning that makes an argument invalid or misleading, regardless of whether the conclusion happens to be true.
Two important points:
A fallacious argument can reach a true conclusion. The conclusion of "All cats can fly; Socrates is a cat; therefore Socrates can fly" is supported by bad reasoning, but a different bad argument might accidentally reach a true conclusion. The fallacy is in the reasoning, not the conclusion.
Pointing out a fallacy doesn't prove the conclusion is wrong. It proves the argument doesn't support the conclusion. The conclusion might still be true for other reasons.
How to Use This Chapter
Don't try to memorize every fallacy. Instead:
- Learn the categories (formal, relevance, ambiguity, statistical)
- Internalize the 10-15 most common ones
- When something "feels off" about an argument, check this reference
- When you spot a fallacy, name it and explain why the reasoning fails
Formal Fallacies
These violate the rules of logical form. The structure itself is broken.
Affirming the Consequent
Form: If P then Q. Q is true. Therefore P is true.
Example:
"If someone is a doctor, they went to university. She went to university. Therefore she's a doctor."
Why it fails: Many things can cause Q. P is just one of them. University leads to many careers.
Response: "That doesn't follow. Many people go to university without becoming doctors."
Denying the Antecedent
Form: If P then Q. P is not true. Therefore Q is not true.
Example:
"If it's raining, the roads are wet. It's not raining. Therefore the roads aren't wet."
Why it fails: Other things (sprinklers, spills, snow melt) can make roads wet.
Response: "The roads could be wet for other reasons."
Undistributed Middle
Form: All A are B. All C are B. Therefore all A are C.
Example:
"All dogs are animals. All cats are animals. Therefore all dogs are cats."
Why it fails: Sharing a category doesn't make two things identical.
Response: "Just because two things share a property doesn't mean they're the same."
Illicit Major / Illicit Minor
Form: All A are B. No C are A. Therefore no C are B.
Example:
"All roses are flowers. No daisies are roses. Therefore no daisies are flowers."
Why it fails: The category B (flowers) is larger than A (roses).
Response: "The broader category includes more than just that subset."
Informal Fallacies of Relevance
The premises may be true but are irrelevant to the conclusion.
Ad Hominem (Attacking the Person)
What it is: Attacking the person making the argument instead of the argument itself.
Variants:
| Type | Example |
|---|---|
| Abusive | "You failed maths, so your budget analysis is wrong" |
| Circumstantial | "You're a landlord, so your housing opinions are biased" |
| Tu quoque (you too) | "You can't tell me to quit smoking. You used to smoke" |
When it IS relevant: If someone's credibility or expertise is directly at issue ("Should we trust this person's testimony?"), their character can be relevant. But it still doesn't address the argument itself.
Response: "Even if that's true about me, does it affect whether my argument is correct?"
Appeal to Authority (Argumentum ad Verecundiam)
What it is: Claiming something is true because an authority says so, without evaluating the reasoning.
When it's fallacious:
- The authority is outside their area of expertise (a physicist commenting on economics)
- Authorities in the field disagree
- The authority has conflicts of interest
When it's legitimate:
- The authority is a genuine expert in the relevant field
- There's consensus among experts
- You're citing expert opinion as evidence, not as proof
Response: "What's their evidence? Being an authority doesn't make them right on this specific point."
Appeal to Popularity (Ad Populum)
What it is: Claiming something is true because many people believe it.
Example:
"Millions of people believe in astrology, so there must be something to it."
Response: "Popularity doesn't determine truth. Millions of people have believed incorrect things throughout history."
Appeal to Tradition
What it is: Claiming something is right because it's always been done that way.
Example:
"We've always used this process. It works fine."
Response: "How long we've done something doesn't tell us if it's the best way. What evidence do we have that it's still effective?"
Appeal to Emotion
What it is: Substituting emotional manipulation for evidence and reasoning.
| Variant | Mechanism | Example |
|---|---|---|
| Appeal to fear | Exploit anxiety | "If we don't act now, disaster will strike" |
| Appeal to pity | Exploit sympathy | "You can't fire me. I have three kids" |
| Appeal to flattery | Exploit vanity | "Smart people like you can see that..." |
| Appeal to anger | Exploit outrage | "Aren't you FURIOUS about this?!" |
Response: "I understand the emotional weight, but what does the evidence actually show?"
Red Herring
What it is: Introducing an irrelevant topic to divert attention from the original issue.
Example:
A: "The company should address the pay gap." B: "We should focus on how much we've grown this year."
Response: "That's a separate issue. Can we get back to the original question?"
Straw Man
What it is: Misrepresenting someone's argument to make it easier to attack.
Example:
A: "We should have more regulation of industrial pollution." B: "My opponent wants the government to control every aspect of business."
How to spot it: Ask "Is that actually what they said?"
Response: "That's not my position. What I actually said was..."
Appeal to Ignorance (Ad Ignorantiam)
What it is: Claiming something is true because it hasn't been proven false (or vice versa).
Example:
"No one has proven aliens DON'T exist, so they probably exist."
Response: "Absence of evidence isn't evidence of absence, but it's not evidence of presence either. The burden of proof lies with the person making the positive claim."
False Dilemma (Black-or-White)
What it is: Presenting only two options when more exist.
Example:
"You're either with us or against us." "Either we ban all guns or we accept mass shootings."
Response: "Those aren't the only two options. What about...?"
Slippery Slope
What it is: Arguing that one step will inevitably lead to extreme consequences, without establishing the chain of causation.
Example:
"If we allow any exceptions to this rule, soon there will be no rules at all."
When it's valid: When there's actual evidence for the chain of events. Some slopes really are slippery.
Response: "What evidence is there that A will actually lead to B, then C, then D? Can you show the mechanism?"
Genetic Fallacy
What it is: Judging something based on its origin rather than its current merit.
Example:
"That idea came from a discredited thinker, so it must be wrong."
Response: "Where an idea comes from doesn't determine whether it's true. Let's evaluate it on its merits."
Begging the Question (Circular Reasoning)
What it is: Using the conclusion as a premise. The argument assumes what it's trying to prove.
Example:
"The Bible is true because it's the word of God, and we know it's the word of God because the Bible says so."
Response: "That's circular. You're assuming the thing you're trying to prove."
Informal Fallacies of Ambiguity
These exploit unclear or shifting meanings of words.
Equivocation
What it is: Using a word with two different meanings in the same argument.
Example:
"The end of a thing is its purpose. Death is the end of life. Therefore death is the purpose of life."
"End" means "goal/purpose" in premise 1 but "termination" in premise 2.
Response: "You're using the word X in two different senses."
Amphiboly
What it is: Ambiguous sentence structure that allows multiple interpretations.
Example:
"I saw the man with the telescope." (Who has the telescope?)
Response: "That statement is ambiguous. Can you clarify what you mean?"
Composition and Division
| Fallacy | Error | Example |
|---|---|---|
| Composition | What's true of parts must be true of the whole | "Every player on this team is great, so the team must be great" |
| Division | What's true of the whole must be true of parts | "This company is profitable, so every division must be profitable" |
Response: "Properties of parts don't necessarily apply to the whole (or vice versa)."
Statistical Fallacies
These misuse data and numbers. See also 05-evidence.md for deeper coverage.
Cherry-Picking
What it is: Selecting only data that supports your conclusion while ignoring contradicting data.
Example: Citing one cold day as evidence against climate change while ignoring decades of temperature data.
Response: "What does the full dataset show, not just selected points?"
Survivorship Bias
What it is: Drawing conclusions only from successes while ignoring failures.
Example:
"Bill Gates dropped out of college and became a billionaire. You don't need college to succeed."
You never hear about the millions who dropped out and didn't become billionaires.
Response: "What happened to all the people who did the same thing but didn't succeed?"
Gambler's Fallacy
What it is: Believing that past random events influence future random events.
Example:
"I've flipped heads 5 times in a row, so tails is 'due'."
Each flip is independent. The coin has no memory.
Response: "Each event is independent. Past outcomes don't change future probabilities."
Base Rate Neglect
What it is: Ignoring the underlying probability when evaluating new evidence.
Example: A medical test is 99% accurate. You test positive. But if the disease affects 1 in 10,000 people, there's still only about a 1% chance you actually have it (because false positives vastly outnumber true positives in a large healthy population).
Response: "What's the base rate? How common is this to begin with?"
Simpson's Paradox
What it is: A trend that appears in several groups reverses when the groups are combined (or vice versa).
Example: Treatment A might look better overall, but Treatment B is actually better for every subgroup. The difference is explained by group sizes and composition.
Response: "Have you looked at the subgroups? Aggregated data can be misleading."
Texas Sharpshooter Fallacy
What it is: Drawing the target after shooting: finding patterns in random data and claiming they're meaningful.
Example: A company notices its best sales quarter coincided with a new logo and claims the logo caused the sales increase.
Response: "Was this pattern predicted in advance, or found after the fact? How many patterns were tested?"
The Fallacy Fallacy
What it is: Assuming that because an argument contains a fallacy, its conclusion must be false.
Example:
A makes a bad argument for climate change. B says: "That argument is fallacious, therefore climate change isn't real."
The fallacy fallacy matters because: Bad arguments can support true conclusions. The existence of a fallacy means the argument doesn't work, not that the conclusion is wrong.
How to Respond to Fallacies
In Conversation
| Situation | Approach |
|---|---|
| Casual discussion | Name the fallacy gently: "I think that might be a false dilemma. Are there other options?" |
| Formal debate | Name it directly and explain why the reasoning fails |
| Someone in power | Frame it as a question: "Could there be other explanations?" |
| Online argument | Often not worth engaging. State it once clearly, then move on |
General Protocol
- Identify the specific fallacy
- Explain why the reasoning doesn't work (don't just name-drop)
- Redirect to the actual evidence or the real argument
- Don't be smug. Fallacy-calling can itself become a rhetorical weapon
When You Catch Yourself
The most important use of fallacy knowledge is self-correction:
- Notice the emotional pull of an argument
- Strip away the rhetoric
- Check: Does the conclusion actually follow from the premises?
- If not, identify why
- Revise your position if needed
Quick Reference: Top 15 Most Common
| Fallacy | One-Line Description | How to Spot It |
|---|---|---|
| Ad hominem | Attacking the person, not the argument | "They said X, but they're a Y" |
| Straw man | Misrepresenting the opponent's position | "So you're saying..." (something they didn't say) |
| False dilemma | Presenting only two options | "Either... or..." with no middle ground |
| Appeal to authority | "Expert X says so" | Check if the expert is relevant |
| Appeal to emotion | Substituting feelings for evidence | Strong emotional language, no data |
| Red herring | Changing the subject | The response doesn't address the original point |
| Slippery slope | One step leads to catastrophe | No evidence for the causal chain |
| Circular reasoning | Conclusion assumes the premise | The "proof" restates the claim differently |
| Appeal to popularity | Many people believe it | Popularity ≠ truth |
| Cherry-picking | Selective data | Ask for the full picture |
| Hasty generalization | Too few examples | Small sample, sweeping conclusion |
| False cause | Correlation treated as causation | "After X, therefore because of X" |
| Tu quoque | "You do it too" | Hypocrisy doesn't invalidate the argument |
| Burden of proof shift | "Prove I'm wrong" | Claimant must provide evidence |
| Equivocation | Shifting word meanings | Same word used differently in the argument |
Key Takeaways
- Fallacies are in the reasoning, not the conclusion. A fallacious argument might still reach a true conclusion
- Learn the categories, then the specific fallacies. Relevance, ambiguity, formal, statistical
- The most common are: ad hominem, straw man, false dilemma, appeal to emotion, and cherry-picking
- Name it and explain it. Just naming a fallacy isn't enough; explain why the reasoning fails
- Apply it to yourself first. You commit fallacies too
- Don't weaponize fallacy knowledge. The goal is truth, not point-scoring
- The fallacy fallacy is real. Bad arguments don't prove the conclusion is wrong