English Toolkit for Thorildsplans Gymnasium

Fallacies

Learn where reasoning goes wrong, and how to repair it.

A fallacy is a problem in the logic of an argument. This guide explains the thirteen fallacies used on the fallacy practice page, grouped by what goes wrong. Each one has an example, an explanation of why it fails, a fairer version of the same point, and a warning about when something that looks like the fallacy is actually reasonable. Try the short practice after each group.

What a fallacy is (and isn’t)

A fallacy is a flaw in the reasoning, not a false claim. An argument can reach a true conclusion for bad reasons. Spotting a fallacy tells you that the argument does not support its conclusion. It does not prove that the conclusion is wrong.

Don’t just shout the name. Calling out “ad hominem!” is not an argument. Explain what the reasoning gets wrong, in your own words. Many arguments that look like fallacies at first glance turn out to be fair, so each fallacy below has a Watch out note for those cases.

The same standards for everyone. Look for fallacies in arguments you agree with, including your own, as carefully as in arguments you disagree with.

1. Attacking the wrong target

These fallacies aim at something other than the actual argument: the person making it, or a distorted version of it.

Ad hominem

Attacking the person who makes an argument instead of the argument itself.

Example

“Why should we listen to Oskar’s ideas about the canteen? He can’t even keep his own locker tidy.”

Why it fails: How tidy Oskar is says nothing about whether his canteen proposal is a good one.

Better version

“Oskar’s plan would cost about 20 kronor more per meal. Where would that money come from?”

Watch out: A person’s interests can matter when the argument depends on trusting them. Pointing out that a study was paid for by the company selling the product is fair: it is a reason to check the evidence more carefully, not a reason to ignore it.

Straw man

Misrepresenting someone’s argument so that it is easier to attack.

Example

A student says, “We should have fewer exams in the same week.” Another replies, “So you want school with no tests at all? Then nobody would learn anything.”

Why it fails: The reply attacks a claim nobody made. Spreading exams out is not the same as getting rid of them.

Better version

“Spreading exams out sounds good, but several courses end at the same time. How would we schedule them?”

Watch out: Summarising someone’s view briefly is fine if the summary is fair. A useful test: would the other person agree that your version describes what they think?

Try it

1. “You can’t trust Noor’s argument about climate policy. She drives a car.” What is the problem?

2. Linnea says the school should serve more vegetarian meals. Ahmed replies, “Linnea wants to ban meat completely.” What is the problem?

3. Fallacy or fair? “The study praising this energy drink was paid for by the company that makes it, so let’s see whether independent studies agree.”

2. Weak evidence

These arguments have evidence, but not enough of it, or not presented honestly enough, to support the conclusion.

Hasty generalisation

Drawing a broad conclusion from too few cases, or from cases that don’t represent the whole group.

Example

“The two Swedish films I’ve seen were slow and depressing. Swedish films are all boring.”

Why it fails: Two films out of thousands tell you almost nothing about Swedish film in general.

Better version

“I didn’t enjoy the two Swedish films I’ve seen, but that’s not enough to judge them all. Can anyone recommend a good one?”

Watch out: A small sample can be enough when the cases are very similar. You only need one spoonful to know whether soup is too salty. The question is whether the sample represents the whole. Personal stories (“My grandad smoked and lived to 94”) are a common form of this fallacy when they are used to outweigh large studies.

Cherry-picking

Selecting only the evidence that supports your side and leaving out the rest.

Example

A politician says, “Unemployment fell in March and in August. Our policies are working.” Unemployment rose in the other ten months.

Why it fails: The full year points the other way. Listeners can’t judge the claim without the missing months.

Better version

“Unemployment rose for most of the year, although it fell in March and August. Here is why we think the trend will change.”

Watch out: Every argument selects some evidence; nobody can include everything. It becomes cherry-picking when the evidence left out would change the conclusion.

Misleading statistics

Using numbers that are accurate but presented in a way that gives a false impression.

Example

A company’s job advert says, “The average salary here is 60,000 kronor a month.” Most staff earn 26,000. The average is pushed up by the owner, who pays himself several million a year.

Why it fails: The number is technically true, but it hides what a typical employee earns. Percentages without the real numbers, graphs with stretched axes, and tiny or unrepresentative samples can mislead in the same way.

Better version

“Most staff here earn around 26,000 kronor a month.” (This is the median: the middle salary when all salaries are put in order.)

Watch out: Statistics are not the problem. Ask what is being counted, what the number is compared with, and what has been left out.

Try it

1. “My cousin used this revision app and got an A, so it’s the best way to study.” What is the main problem?

2. Which of these is an example of cherry-picking?

3. “9 out of 10 students prefer our study app!” The company asked ten students who already pay for the app. What is the problem?

3. Faulty cause and effect

These arguments claim that one thing causes another, or will lead to another, without showing how.

False causation

Assuming that one thing caused another just because they happened together or one after the other.

Example

“Students who eat breakfast get higher grades, so eating breakfast makes you smarter.”

Why it fails: The two things are linked (a correlation), but that doesn’t show that one causes the other. Both could depend on something else, such as regular sleep or home routines.

Better version

“Breakfast and grades are linked. A study that takes sleep and home life into account could show whether breakfast itself makes a difference.”

Watch out: A link between two things is often the first clue to a real cause. It is a reason to investigate, not a reason to conclude.

Unsupported slippery slope

Claiming that one step will lead to a chain of extreme consequences, without evidence for each step.

Example

“If we let students use phones at lunch, soon they’ll use them in lessons, then nobody will listen to the teacher, and grades will collapse.”

Why it fails: Each step is simply asserted. Nothing shows that lunchtime phone use leads to the next step, let alone to the last one.

Better version

“Schools that allowed phones at lunch reported more phone use in lessons. Let’s look at what happened to their results.”

Watch out: Some chains of consequences are real. That is why this fallacy is unsupported slippery slope: it is only a fallacy when the steps are claimed rather than shown.

Try it

1. “Since the new head teacher started, the canteen food has got worse. She must have changed the menu.” What is the problem?

2. Fallacy or fair? “If we raise the bus fare, some students may stop taking the bus. In a survey, 30% said they could not afford a higher fare.”

4. Pressure instead of reasons

These arguments try to make you agree through fear or a famous name rather than through evidence.

Appeal to fear

Using fear to persuade instead of giving evidence that the claim is true.

Example

An advert for a home alarm shows a frightened child and a dark window. It gives no information about how well the alarm works.

Why it fails: Fear makes you want to buy something, but it tells you nothing about whether this product protects your home.

Better version

“Our alarm alerts a security centre within ten seconds. Independent tests rated it the most reliable of the brands they checked.”

Watch out: Some fears are justified. A warning about a real danger, backed by evidence, is not a fallacy. It becomes one when the fear replaces the reasons.

Uncritical reliance on authority

Accepting a claim because of who said it, when that person is not an expert on the subject or the evidence points the other way.

Example

“A famous footballer says this supplement boosts your immune system, so it must work.”

Why it fails: Being an expert at football doesn’t make someone an expert on medicine. They may also be paid to say it.

Better version

“What do doctors and immunology research say about this supplement?”

Watch out: Relying on real experts speaking about their own field, such as a virologist on viruses, is good reasoning, especially when most experts agree. The problem is relying on authority uncritically.

Try it

1. “Buy our antivirus software today, or hackers will empty your bank account tonight!” The advert gives no other information. What is the problem?

2. Which of these is a reasonable use of authority?

5. Rigged choices and slippery words

These arguments trap you with the way the question is set up, or with the way a word is used.

False dilemma

Presenting only two options when more exist.

Example

“Either we build the new sports hall, or we’re giving up on students’ health.”

Why it fails: There are many other ways to support students’ health, and someone can care about it and still think the sports hall is the wrong choice.

Better version

“I think a new sports hall is the best way to get students more active. If you disagree, what would you do instead?”

Watch out: Sometimes there really are only two options. A referendum may only offer yes or no.

Equivocation

Switching between two meanings of the same word in the middle of an argument.

Example

“Evolution is just a theory. Theories are only guesses, so evolution is only a guess.”

Why it fails: In everyday English, a theory can mean a guess. In science, a theory is a well-tested explanation supported by a large body of evidence. The argument slides from one meaning to the other.

Better version

“Evolution is a scientific theory. What is the evidence for it, and how has it been tested?”

Watch out: Many words have more than one meaning, and that’s fine. The problem is only when the meaning changes partway through and the conclusion depends on the switch.

Try it

1. “Either we ban phones in school completely, or students will never be able to concentrate.” What is the problem?

2. “Nothing is better than a good education. A cheese sandwich is better than nothing. So a cheese sandwich is better than a good education.” What is the problem?

6. Arguments that can’t lose

These arguments are built so that nothing could ever count against them. That can make them feel strong, but it means they never really prove anything.

Begging the question

Using the conclusion as one of the reasons for it. This is also called circular reasoning.

Example

“This news site is trustworthy because it says it only publishes accurate news.”

Why it fails: The argument assumes the site is trustworthy in order to prove that it is trustworthy.

Better version

“This news site corrects its mistakes publicly, and independent fact-checkers rate its reporting as accurate.”

Watch out: In everyday English, “that begs the question” often means “that raises the question”. In logic, it names this fallacy. Restating a claim to make it clearer is fine; it becomes circular when the restatement is offered as proof.

Unfalsifiable claims

Making a claim that no possible evidence could ever show to be wrong.

Example

“My lucky stone protects me. When something bad happens, that just shows how much worse it would have been without the stone.”

Why it fails: Good outcomes count as proof, and bad outcomes count as proof too. If nothing could count against a claim, no evidence can really support it either.

Better version

“I feel calmer when I carry my lucky stone.” (This is a claim about your own feelings, and it doesn’t pretend to be more.)

Watch out: Some claims are hard to test today but could be tested in principle. Others, such as values or personal preferences, are not meant to be tested at all. The problem is presenting a claim as a proven fact while ruling out any evidence against it.

Try it

1. “Smoking is bad for you because it harms your health.” What is the problem?

2. “Ghosts are real. The fact that nobody has ever recorded one just shows how good they are at hiding.” What is the problem?

Using this with sources

Source evaluation and fallacy spotting work together. First ask whether a source is reliable, using the questions in the source evaluation guide. Then ask whether its reasoning holds up. A reliable source can still argue badly, and a weak source can occasionally make a fair point.

Check your own reasoning

Before you hand in an argumentative text, read it as if someone you disagree with had written it.

If you find a fallacy, repair it. The Better version examples above show how. When you are ready, try the fallacy flashcards and matching game.

Your answers are checked in your browser and are not sent to your teacher.