The short version
Somebody proposes A. The objection isn’t to A — it’s to E, which A is said to lead to by way of B, C and D.
Those middle steps are the entire argument, and they are the part that never gets made. The proposal is answered as though it were the destination.
Where the name comes from
The older names are better pictures. The thin end of the wedge — the narrow edge that lets the rest be driven in behind it. The camel’s nose — let the camel get its nose into the tent and in due course you have a camel in the tent and you’re outside it, in the weather.
Both capture the same claim, which is a real one worth taking seriously: that a small conceded thing can create the conditions for a large unconceded one. The fallacy isn’t in believing that ever happens. It’s in asserting it instead of showing it.
What it looks like
Let people work from home on Fridays.
Then it’s Thursdays, then nobody comes in at all, then why are we paying for an office.
Log who accesses the production database.
Then it’s keystrokes, then it’s screen recording, then we’re all on a monitoring dashboard.
Add one exception to the returns policy for damaged goods.
Then we’ll be refunding everyone for anything, forever.
Why it doesn’t work
A chain is no stronger than its weakest link — and this move never shows you the links. It shows you the two ends and invites you to fill in the middle yourself, which you will, because your imagination is obliging and nobody has been asked to defend anything. Each step needs its own argument, and the whole design of the move is that none of them gets one.
Then there’s the arithmetic, which almost never gets done. Suppose each step really is 80 per cent likely to follow from the one before — generous, for a speculative chain nobody has evidenced. Four steps is 0.8 × 0.8 × 0.8 × 0.8, which is about 41 per cent. Chains of individually-plausible steps become improbable startlingly fast, and intuition doesn’t track it at all, because you assess the steps one at a time and each one sounds reasonable on its own.
And there’s what the move buys, which is the same thing a straw man buys by a different route. It lets you argue against something nobody proposed. You get to fight the most alarming version of your opponent’s position that you can imagine, rather than the one actually on the table — a straw man built out of the future instead of out of a misquotation.
This is not a slippery slope
- A precedent argument with a stated mechanism. “If we grant this exception, we’ll have no principled basis for refusing the next one” is a real argument, and often a strong one, because it says why the second step follows. That’s checkable, and therefore disputable.
- Slopes that genuinely are slippery. Some are. Institutions really do erode, exceptions really do become rules, and pretending otherwise is its own failure. The question is never whether slopes exist — it’s whether this one has been shown.
- Empirical prediction. “Three other places did this; here’s what followed” is evidence. It may be bad evidence or the cases may not compare, but it is the right kind of thing, and it can be argued with.
- Asking where the line is. “What wouldn’t you allow?” is a fair question, and a proposal whose answer is “nothing in particular” has a real problem. Asking for the limiting principle is legitimate. Asserting there can’t be one is the move.
The test is about the middle rather than the ends: can they walk you from here to there one step at a time, saying why each step follows from the last? If they can, it’s an argument, and you should go and find the weakest link. If they jump straight to the end, there was never a chain — only two points and an implication doing unpaid work.
How to call it out
- “Walk me from here to there. What’s the step immediately after this one, and why does it follow?”
- “Which step do you actually think is the likely one? Let’s argue about that.”
- “Is there anything between this and that which we could agree to stop at?”
That last question is usually the productive one, because it converts a fight about the destination into a negotiation about a limiting principle — and a limiting principle is the thing that makes a slope not slippery. Ask it honestly and one of two useful things happens. Either you find the stopping point, write it down, and the objection dissolves because it has been answered rather than dismissed. Or it turns out that neither of you can name one, in which case the worry was legitimate the whole time, and you’ve just learned the most important thing about your own proposal.