The missing bridge

If bananas are hats, then apples are cats.

Read as English, that sentence is nonsense; run through classical propositional logic, it comes out true, and both verdicts are correct. The trouble starts when if wanders between the two systems without anyone saying what kind of connection it's supposed to carry.

Only one row loses#

Let P mean “bananas are hats” and Q mean “apples are cats.” The material conditional P → Q is false in exactly one case: P is true and Q is false. Every other combination counts as true.

The arrow doesn't inspect fruit or plausibility or anything else about the world. All it asks is whether the conditional has been broken — did bananas turn out to be hats while apples failed to be cats? If not, the formula survives.

In the ordinary world both claims are false, so we land on the last row and the conditional comes out true. You get the same answer if you rewrite P → Q as ¬P ∨ Q: bananas aren't hats, so the left side of that disjunction settles the matter before apples ever come up. Logicians call this vacuous truth.

That answer sounds like a lawyer's trick, and the reason is that ordinary speech doesn't hear a four-row operator. It hears an invitation: imagine bananas being hats, then tell me what happens to apples. And nothing happens. Once you're in that imagined world, the sentence hands you no rule connecting the categories, no cause, not even a hint about where to look for one.

English expects a hinge#

Philosophers still argue about conditionals, largely because the familiar little word is doing several jobs at once. The Stanford Encyclopedia of Philosophy describes the truth-functional account as logic's first surprise: elegant, useful for mathematics, and an awkward fit for a lot of what we say about the world.

“If the build fails, the deploy is blocked” states a rule somebody configured. “If the glass falls, it breaks” is a claim about physics, while “if she left at six, she's home by now” is offering evidence, and “if I had left at six, I'd be home” asks you to consider a world nearby to this one. Nearly the same grammar every time, resting on a different kind of connection underneath.

Classical logic flattens all of that on purpose. If you want a connective you can calculate with, its behavior has to depend on truth values alone — not physics, not what the speaker was getting at. The price is that the arrow becomes a poor translator for any English sentence whose whole point is the relationship.

Lewis Carroll stages a related mistake in his 1895 dialogue “What the Tortoise Said to Achilles”. The Tortoise accepts the premises and every new hypothetical Achilles writes down, but refuses to take the inferential step, so Achilles keeps adding statements as though one more sentence could stand in for knowing how to use a rule. It can't, and the regress never ends.

The banana sentence trips over the opposite error: you can apply the formal rule perfectly and still have no reason on the other side. So when someone hands me an if-then claim, I've started asking what's supposed to make the consequent follow. A truth table? A causal story, a contract, something they watched happen? Usually there's an answer. Sometimes there isn't.

That's why the nonsense version earns its keep. Respectable nouns hide the gap — “if users care, they'll click” sounds connected well before anyone produces evidence. Bananas and hats extend no such courtesy. A conditional can be true and still have nothing to say.