The short version
Somebody presents two options as if they were the whole list. From then on, any objection to one counts as support for the other.
The trick isn’t in either option — both are usually real, and one of them might even be right. It’s in the word either.
What it looks like
- Rewrite the whole system
- Live with the bugs
- Fix the three modules that cause most of them
- Rewrite one subsystem at a time
- Ship a workaround now and decide next quarter
- Decide it isn’t worth fixing, and say so out loud
- You’re with us
- You’re against us
- Agree with the goal, object to the method
- Agree with most of it, not this part
- Not have a stake in it at all
- Not know enough yet to have a side
Why it doesn’t work
Here’s the part worth understanding: if the two options really were the whole list, the reasoning would be perfectly valid. “Either A or B; not A; therefore B” is textbook-correct logic. Nothing wrong with it.
Which means the entire argument is won or lost on whether the list is complete — and that question gets settled before anybody starts reasoning. By the time you’re choosing, you’ve already accepted the menu. That’s why this works on smart people. They’re busy evaluating the two options carefully, which is precisely the wrong thing to be doing.
The second cost is that it punishes precision. Anybody who says “some of A, some of B, depending on the case” gets heard as evasive, indecisive, or a fence-sitter. Which is often the most accurate position available.
This is not a false dichotomy
- Real binaries. Some things genuinely have two values. You ship or you don’t. The flag is on or off. She’s on the flight or she isn’t. Calling a real binary a false dichotomy is its own mistake, and an irritating one.
- Narrowing under a real constraint. “We’ve got two weeks and two people, so realistically it’s A or B” is a constraint, not a fallacy. Constraints are legitimate — as long as they’re stated, and as long as you’re allowed to question them.
- Being made to decide. Somebody saying “pick one, we’re out of time” isn’t committing a fallacy. That’s just management.
The test: name a specific third option and see whether it survives contact. If you can name one and it holds up, the list wasn’t complete. If every third option you come up with is worse than both of the originals, then the dichotomy was probably real and you should pick one.
How to call it out
- “Third option: [name it]. What’s wrong with that one?” Always propose something specific — “there might be other options” gets waved off in half a second.
- “Are those the only two there are, or the only two we’ve thought of?”
- “I’m rejecting A. That isn’t the same thing as choosing B, and I’d like that on the record.”
That last distinction is the one that matters most. Most of the damage from a false dichotomy isn’t the bad choice — it’s that turning down one option gets logged as endorsing the other.