Straw men, slippery slopes and other tricks that sound like arguments

You have been in that conversation. Something felt wrong about what the other person said, but you could not put your finger on what – and by the time you worked it out in the car home, it was too late.

This episode is about being able to put your finger on it straight away.

What a fallacy actually is

A fallacy is an error in the argument, not in the conclusion. It is a piece of reasoning where the premises do not carry the claim, even though it may sound entirely right.

That is an important distinction, and it is the first thing people get wrong. Someone arguing terribly that the earth is round does not make the earth flat. We will come back to that.

Traditionally a line is drawn between formal fallacies, where the structure of the logic itself breaks, and informal ones, where the problem lies in the content, the language or the context. The informal ones are what you meet in practice, and they are what we spend the time on.

The straw man

The most common trick of all. You build a weaker version of what the other person said, knock it down, and act as though you have won.

— I think the speed limit should be lowered in residential areas. — So you want everyone crawling along everywhere?

That is not what was said. And here is the point that often gets missed: the straw man is usually not bad faith. We hear what we expect to hear, and build a version of the other side that is easier to disagree with.

The straw man, side by side

What was said

Lower the limit in residential streets

• Applies to one particular place
• Says nothing about other roads
• No figure attached

What was answered

Walking pace everywhere

• Applies to every road
• An extreme nobody proposed
• Easy to dismiss

The counter-move is the steelman: repeat the other person’s position in the strongest version you can manage, and ask them to confirm you have understood it. Do that, and you quite often find the disagreement is smaller than you thought.

Ad hominem – and when it is not a fallacy

Attacking the person instead of the argument:

— You can’t have a view on money, you went bankrupt yourself.

But you have to be precise here, because ad hominem is not always a fallacy. It is one when the trait is irrelevant to the claim. When it is relevant, it is a legitimate point:

• “He is wrong about climate because he is a bad father” – a fallacy.
• “The report was paid for by the industry it assesses” – not a fallacy. That is relevant information about credibility, and something any journalist is expected to disclose.

When the person is relevant, and when they are not

Claim about the person	Does it bear on the matter?	Verdict
“He is a bad father”	No. Fatherhood says nothing about climate science.	Fallacy
“The report was paid for by the industry it assesses”	Yes. A conflict of interest affects how much weight an assessment deserves.	Valid point
“She has never worked in the field”	It depends. Relevant for experience-based knowledge, less so for a calculation.	To be judged

The line is whether the trait has anything to do with the matter at hand.

The appeal to authority – which is usually fine

This is the most misused label on the list. People shout “appeal to authority” whenever anyone cites an expert, and that is wrong.

Relying on relevant experts is normal and sensible thinking. You do not calculate the load-bearing capacity of the bridge yourself. The fallacy arises in three cases:

• Wrong field. A Nobel laureate in physics holding forth on nutrition.
• No real authority. Titles that sound weighty without meaning anything.
• No such agreement. One researcher presented as “the research shows”, when the field is divided.

Three questions before trusting an authority

1. Is the person an expert on this exact thing?Not on something nearby – on this particular claim.
2. Is the authority real?What sits behind the title? Training, publications, practice?
3. Does the field agree?One researcher is not “the research”. Is this the breadth, or one extreme?
4. Three yeses is good reason to trust itOne no does not make the claim false – only that the authority does not settle it.

Rule of thumb: check whether the person is an expert on this specific thing, and whether they represent the breadth of the field or one extreme of it.

The false dilemma

Two options presented as though they were the only ones.

— Either we cut welfare, or the country goes bankrupt.

There is almost always a third, fourth and fifth possibility. The trick works because two choices are easy to hold in your head, and because it pushes you to pick a side before you have had a chance to ask whether there are more.

The slippery slope – which often is not a fallacy

— Allow this, and we will end up allowing anything.

Here is the nuance that is usually missing: a slippery slope is only a fallacy when the causal chain is not argued for. If you can show that each step actually follows from the previous one, it is a perfectly valid argument – and that kind of reasoning is used daily in law, in safety work and in politics.

Two slippery slopes, the same shape

The links hold

1. AThe step is argued for
2. BThe step is argued for
3. CValid argument

The links do not hold

1. AThe step is argued for
2. ?No argument given
3. ZFallacy

The question is never “are they using a slippery slope?”, but “do the links hold?”.

The circular argument

The conclusion is already contained in the premise.

— Why should we trust him? Because he is trustworthy.

Longer versions are harder to spot, because the circle can run through several paragraphs before it returns to where it started.

Shifting the burden of proof

— Can you prove it doesn’t exist?

The burden of proof sits with whoever makes the claim. You owe nobody a proof that something does not exist – and in most cases it is not possible anyway.

Whataboutism

Answering criticism by pointing at something else.

— You delivered late. — Yes, but you did the same last year.

That may well be true. It is simply not an answer. Notice that it often works: the conversation moves, and the original question is never addressed.

No true Scotsman

A counter-example is defined away instead of being taken seriously.

— No Norwegian skips the Sunday walk. — My neighbour does. — Then he is no true Norwegian.

The rule is rescued by changing what it is about, mid-argument.

Appeals to nature and to tradition

— It is natural, therefore it is good. — We have always done it this way.

That something is natural says nothing about whether it is good – arsenic is natural, and vaccines are not. That something is old says nothing about whether it works. Both can be useful indications, but neither is an argument.

Hasty generalisation, and correlation that is not cause

Two that belong together.

Hasty generalisation is drawing a rule from too few cases: two bad experiences with a brand, and the brand is bad. Correlation is not causation is when two things track each other without either causing the other. Ice cream sales and drownings track each other closely. The cause is summer.

The third variant is after, therefore because of: something happened after something else, so the first must have caused the second.

The Gish gallop and motte-and-bailey

Two tricks you notice most in online debate.

The Gish gallop is firing off so many claims in so little time that the other side cannot possibly answer them. It takes twenty seconds to assert something and ten minutes to refute it, and that asymmetry can be exploited deliberately.

Motte-and-bailey is alternating between two versions of the same position: a bold one that is hard to defend (the bailey), and a cautious one almost nobody disagrees with (the motte). Attack the bold version, and the arguer retreats to the cautious one – and once the attack is over, they are back in the bailey.

The fallacy fallacy

The most important one, and the one that makes this whole list dangerous to learn.

A bad argument does not make the conclusion false. If someone uses a straw man to defend a position, the position may still be right – they just have not shown it yet.

A bad argument for something true is still a bad argument – and the claim is still true.

Example: “The Earth is round because my grandmother said so.” The argument does not hold. The Earth is round anyway.

An honest warning: people who have just learned about fallacies often become unbearable. They start shouting Latin names instead of answering the point, and use the list as a weapon rather than a tool. That is using the knowledge backwards.

The most useful thing is not catching other people. It is noticing when you are doing it yourself – because a straw man is far easier to build than to spot, and nearly all of us build them without meaning anything by it.

How to actually respond

Four moves that work better than quoting Latin:

1. Repeat the position back. “As I understand you, you think …” If it was a straw man, it is cleared up immediately.
2. Ask for the missing link. “How does the second thing follow from the first?” Far more effective than giving the trick a name.
3. Hold on to one question. Against whataboutism and the gallop: “We can come to that – first I would like an answer to this.”
4. Separate the matter from the person. Including when you would rather not.

Sources are given in the episode.

Main topics

• What a fallacy actually is, and what it is not
• The straw man – answering something the other person did not say
• Ad hominem, and when it is actually relevant
• The false dilemma
• The slippery slope, and when it is not a fallacy
• The appeal to authority, and why it is usually fine
• Circular arguments and shifting the burden of proof
• Whataboutism
• No true Scotsman
• Appeals to nature and to tradition
• Hasty generalisation, and correlation that is not cause
• The Gish gallop and motte-and-bailey
• The fallacy fallacy

Questions this episode answers

• What is a logical fallacy?
• What is a straw man?
• Is ad hominem always a fallacy?
• When is it acceptable to cite an authority?
• What is a false dilemma?
• Is the slippery slope always wrong?
• What is whataboutism?
• Who carries the burden of proof in a disagreement?
• Does a bad argument make the conclusion false?
• How do you answer a rhetorical trick without becoming insufferable?

Key points

• A fallacy is an error in the argument, not in the conclusion.
• The straw man is the most common trick, and the easiest to commit without meaning any harm.
• Ad hominem and the appeal to authority are only fallacies when the trait is irrelevant to the question.
• A slippery slope with a real causal chain is a perfectly valid argument.
• Whoever makes the claim carries the burden of proof. You do not have to prove a negative.
• Someone arguing badly does not mean they are wrong. That is a fallacy of its own.

Sources and references

• Douglas Walton – Informal Logic (2nd ed. 2008)