Logical fallacies: the fifteen a student meets, each with a sentence that commits it and the honest version

A logical fallacy is an argument whose conclusion does not follow from its reasons, even when the conclusion happens to be true. The fifteen below are the ones that turn up in school essays and debates: attacking the person, misstating the other side, offering two options when there are more, and so on. Each is shown twice: a sentence that commits it, and the honest version that argues the same point with the evidence it would need. Then the paragraph that answers an objection without any of them.

What a logical fallacy is, and what it is not

An argument has premises (the reasons given) and a conclusion (what they are supposed to show). A fallacy is a gap between the two. Two kinds matter here. In a fallacy of form, the step itself is broken: the conclusion would not follow even if every premise were true. In a fallacy of relevance, the premise is beside the point: true, perhaps, but about something else.

What a fallacy is not: proof that the conclusion is false. A later bell may be a good idea even if someone argued for it badly. Catching a fallacy tells you that sentence did not do its job; it tells you nothing about the world. That is why the last section of this page is about the mistake good students make once they have learned the list.

The modern list has two books behind it. Hamblin (1970) traced the standard fallacies back through two thousand years of logic textbooks and found most of them poorly defined. Walton (2008) rebuilt them as moves in a dialogue: a fallacy is a move that looks like it answers the other side and does not. That is why every item below has an honest version, the move that actually answers. In the Academic Writing course this is idea 13, logical fallacies, between counter-argument and rebuttal and essay structure.

The fifteen, each committed and then fixed

Every sentence is ours, about the running example of this section (a proposal to move a school's first bell to 9 am) and other school-sized topics. No politician, advertiser or author is quoted. The topics are deliberately small; the forms are the point.

1. Ad hominem (against the person)

Attacking the arguer instead of the argument.

Committed: Ms. Reyes wants a later bell because she likes sleeping in.

Honest: The proposal's attendance figures come from one term; they need a second term before the board acts.

What it would have to show: a weakness in the evidence, not in the person who presented it.

2. Straw man

Answering a weaker claim than the one the other side made.

Committed: Supporters want school to start at noon.

Honest: Supporters want the first bell at 9 am, one hour later. The question is whether one hour buys the attendance they claim.

What it would have to show: the real claim, restated in words the other side would accept, then answered.

3. False dilemma

Offering two options when there are more.

Committed: Either we move the bell or we accept failing students.

Honest: Moving the bell is one option. A later first period for the senior grades alone and a sleep unit in health class are two more, and each should be costed.

What it would have to show: that the other options were considered and ruled out.

4. Slippery slope

Claiming one step leads to an extreme with no reason it would.

Committed: Move the bell to 9 and soon students will demand no school at all.

Honest: The step proposed is one hour, and it carries one cost, the afternoon. Argue about that step.

What it would have to show: a mechanism by which the first step causes the last.

5. Hasty generalisation

Treating one case as a rule.

Committed: My cousin's school tried it and it was chaos, so it never works.

Honest: One school's experience is one data point. How many schools have tried it, and what happened at each?

What it would have to show: enough cases, chosen fairly, to support "never".

6. Appeal to popularity

Treating what most people want as proof that it is right.

Committed: Most students want it, so it must be right.

Honest: What students want is one input. The question is what the change does to attendance and to the bus budget.

What it would have to show: evidence about the outcome, not the vote.

7. Appeal to irrelevant authority

Citing someone famous outside their field.

Committed: A famous actor said teenagers need more sleep.

Honest: The researchers who have measured adolescent sleep found most teenagers short of it on school nights; cite them, by what they measured.

What it would have to show: an authority on this question, and what their evidence was. Every statistic in the essay rests on a source it names; citing what the statistic rests on shows how.

8. Appeal to emotion

Letting a feeling stand in for the reason.

Committed: Think of the exhausted child crying on the bus.

Honest: Half of the students surveyed report under seven hours of sleep on school nights; the child on the bus is what that number looks like.

What it would have to show: the fact the feeling points at. The feeling itself is not a fault, as the three appeals explains; a feeling asked to do a reason's job is.

9. Circular reasoning (begging the question)

Giving the claim as its own reason.

Committed: A later bell is better because starting later is an improvement.

Honest: A later bell is better because first-period attendance rose in the pilot and no other change the board tried moved it.

What it would have to show: a reason that is not the claim in other words. A thesis that does this fails the second question of the thesis test.

10. Post hoc (post hoc ergo propter hoc, "after this, therefore because of this")

Treating what came after as what was caused.

Committed: Attendance rose the year after the pilot, so the pilot caused it.

Honest: Attendance rose the year after the pilot. What else changed that year, and what happened at a comparable school that kept the early bell?

What it would have to show: that the other causes were ruled out, usually by a comparison.

11. False analogy

Comparing two things that differ in the way that matters.

Committed: A school is a factory; you do not move a factory's shift for the workers' comfort.

Honest: The analogy breaks where it matters: students are not paid to be awake, and the product of a school is their attention.

What it would have to show: that the two things are alike in the respect the argument needs.

12. Red herring

Changing the subject.

Committed: Before we discuss the bell, why has the board not fixed the cafeteria?

Honest: The cafeteria is a separate motion. On the bell, the one question is the attendance gain against the transport cost.

What it would have to show: that the new subject bears on the old one. It does not, so it goes.

13. Tu quoque ("you too")

Dismissing a claim because the claimant does not live by it.

Committed: You cannot argue for more sleep; you stay up gaming.

Honest: The arguer's habits do not bear on whether the claim is true. Does the evidence say teenagers are short of sleep?

What it would have to show: a fault in the claim, not in the claimant's bedtime.

14. Loaded question

Building the accusation into the question.

Committed: Why does the board hate teenagers?

Honest: Why has the board kept the 8 am bell? The answer may be bus contracts, after-school care or habit, and each can be examined.

What it would have to show: the question asked without the verdict inside it.

15. Appeal to tradition

Treating the age of a practice as a reason for it.

Committed: School has always started at 8, and that is how it should stay.

Honest: The 8 am bell was set around bus contracts and farm hours. Ask what it was for and whether those reasons still hold.

What it would have to show: that the original reason still applies.

How to find them in your own draft

1. Read each paragraph and write its argument in the margin as "because A, therefore B". If you cannot, the paragraph has no argument yet.
2. Ask whether B follows from A alone. If it needs a hidden step, write the step down and check it.
3. Underline every "everyone", "always", "either", "obviously", "clearly". Each is a place where a generalisation, a dilemma or an unargued claim likes to hide.
4. Check that every attack is on a claim, not a person.
5. Check that every statistic has a source and a comparison. A number with nothing beside it proves little.

The paragraph-level version of this check, the topic sentence and the sentence that says what the evidence shows, is the paragraph checklist. Finding the one line that does not follow is chapter 4 of the free playbook.

The counterargument paragraph, written honestly

The Academic Writing course's idea 12, counter-argument and rebuttal, is the paragraph where fallacies are most tempting, because it is the one place you have to state a view you do not hold. Three moves, in order:

1. State the strongest objection fairly, in words the other side would accept. This is the opposite of a straw man.
2. Concede what is true in it. To concede is to admit the point; a reader trusts a writer who can.
3. Rebut (answer it) with evidence, or narrow your claim so the objection no longer reaches it.

Our paragraph, with the moves marked in brackets:

The strongest case against a 9 am bell is the afternoon. Practices, part-time jobs and younger siblings' pickups are all built around a 3 pm finish, and an hour added to the morning is an hour taken from them [objection]. That cost is real, and it falls hardest on students who work; no survey figure makes it go away [concession]. But it is a cost to the senior grades, who hold the jobs, and the attendance gain is also theirs. A pilot that moves only Grades 11 and 12, and lets them choose a last-period study block, keeps the afternoon for the students who need it while testing the claim where it matters [rebuttal, and the claim narrowed].

The same paragraph done badly:

Some people say moving the bell will ruin after-school life. But students can adjust, and the ones complaining mostly just want to leave early. The benefits clearly outweigh any minor inconvenience.

A straw man ("ruin after-school life" is not what anyone said), an ad hominem ("just want to leave early"), no concession, and "clearly" doing the work the evidence should. The two paragraphs argue for the same thing; only one of them would persuade a board.

Nussbaum and Schraw (2007) studied students writing argumentative essays and found that those who weighed the counter-arguments against their own produced more integrated arguments than those who argued one side. Kuhn (1991) found that most people, asked to argue a view they held, struggled to generate the strongest case against it at all. Both findings say the same thing: the counter-argument paragraph is the hard one, and it is where the marks are.

Where the counterargument goes in the essay

In most school essays, after the body paragraphs and before the conclusion: the reader has heard your reasons and now sees you can answer the best objection to them. Sometimes it goes first, when the objection is what everyone already believes (if your readers all assume the bell cannot move because of buses, clear that away before you argue anything). That is a structural choice, the course's idea 14, and the hub where the counterargument goes in the essay places it among the other parts.

The one that catches good students: the fallacy fallacy

Once you can name fifteen fallacies, every opponent's sentence looks like one, and it is tempting to write "this is a slippery slope" and move on as if you had won. You have not. The other side's conclusion may still be true, and a marker knows that a label is not a rebuttal.

The honest move is the hard one: rebuild their argument in its strongest form, the version a careful person on that side would actually make, and answer that. If their sentence committed a hasty generalisation, ask what the evidence would be if they had it, and deal with that evidence. An essay that beats the strongest version of the other side is persuasive. An essay that collects its opponent's slips is a list.

The fifteen on one screen

Fallacy	The tell
Ad hominem	The person is attacked, not the claim
Straw man	The other side's claim is weaker than the one they made
False dilemma	Two options, when there are more
Slippery slope	A chain to an extreme with no mechanism
Hasty generalisation	One case, treated as a rule
Appeal to popularity	Many people want it, so it is right
Irrelevant authority	Famous, but not on this question
Appeal to emotion	The feeling stands in for the reason
Circular reasoning	The claim is its own reason
Post hoc	After it, therefore because of it
False analogy	Alike, but not where it matters
Red herring	The subject has changed
Tu quoque	The arguer does not live by it
Loaded question	The accusation is inside the question
Appeal to tradition	It is old, so it is right

The fix for every one is the same shape: restate the point, then supply the evidence it would need.