Here's a thought experiment that splits a dinner table down the middle. A comedian gets up and says: a church just raffled off a Hummer to bring in new members. Half the room laughs out loud. The other half goes quiet and a little stiff. Same words. Same room. Same night. One group is delighted, the other is mildly appalled. Nobody's wrong, and nobody's lying about their reaction.
That gap — the laugh and the wince landing on the exact same sentence — is the puzzle this whole section is built around. And it has an answer, worked out in a psychology lab, that turns "is this funny?" from a coin flip into something close to a recipe.
The person who cracked it is Peter McGraw, a behavioral scientist who runs something called the Humor Research Lab at the University of Colorado — and yes, it's deliberately abbreviated HuRL, because comedy people can't help themselves. Working with his collaborator Caleb Warren, McGraw published a paper in 2010 in the journal Psychological Science with a title that tells you exactly what they were after: "Benign violations: making immoral behavior funny." Their claim is bold and narrow at the same time. Humor happens when, and only when, three things are true at once.
Here's the first one. There has to be a violation. Something has to be wrong. McGraw defines a violation as anything that threatens the way you believe the world ought to be. That sounds heavy, but it covers an enormous range. A pun violates the rules of language. A play-fight violates your body's sense of safety. An insult violates social norms. A non-sequitur violates logic. The church raffling off a Hummer violates a vague sense that sacred things and gas-guzzling SUVs don't belong in the same sentence. Strip the fancy word away, and "violation" just means: something here is off.
Now hold that, because on its own it's useless. Most wrong things aren't funny at all. A car crash is a violation. A tax audit is a violation. Stepping in something on the sidewalk is a violation. None of those make you laugh — they make you upset, or scared, or grossed out. So the violation is necessary, but it's nowhere near enough.
That brings the second condition. The violation also has to feel benign. Safe. Okay. Acceptable. Somehow your brain has to register "this is wrong" and "this is fine" in the same instant. McGraw and Warren built on earlier work by a linguist named Tom Veatch, and the example they keep coming back to is the oldest joke in the primate playbook: tickling. Think about what tickling actually is. It's a mock attack. Someone's fingers are darting at the soft, vulnerable parts of your body — your ribs, your neck — the exact spots a predator would go for. Your nervous system reads "threat." And yet you laugh, because at the same moment you know it's your friend, it's a game, nobody's getting hurt. Threatening and harmless, simultaneously. That's the whole theory in one squirming, giggling package.
So here's the line worth carrying out of this section: a joke is a tickle made out of words. It pokes at something that feels a little dangerous, in a way your brain knows is safe — and the laugh is your body's all-clear signal.
And that explains why so many things that should be funny just sit there, dead. The theory's real power isn't that it tells you when things are funny. It's that it predicts, precisely, the three different ways a joke can die. This is the part that turns it from an interesting idea into a working tool.
The first failure mode is too benign. There's a violation, technically, but it's so weak, so safe, so toothless that there's nothing to resolve. The tickling experiment makes this concrete. Stop the mock attack — just rest your hand on someone's arm and don't move — and the laughter stops instantly. No threat, no laugh. In conversation, this is the joke that's so inoffensive it isn't a joke at all. It's the corporate pun on a billboard. It's the dad joke that lands with a groan precisely because there was never any edge to it. Pure benign with no violation isn't comedy. It's just a pleasant, forgettable noise.
The second failure mode is the opposite, and it's the one that ends careers. Too threatening. The violation is real, but nothing makes it feel safe. Push the tickling too hard, get too aggressive, and the laughter flips to panic or anger in a heartbeat. McGraw's lab calls this a malign violation — the wrongness with no off-ramp. This is the joke at a funeral that's too soon and too raw. It's the remark that punches at something the listener actually cares about, with no signal that it's okay. The audience doesn't think "that's edgy." They think "that's cruel," and they're done with you.
The third failure mode is the sneaky one, and it's the reason a lot of decent jokes still flop. The two perceptions have to happen simultaneously. Not the violation, then a beat, then the safety. Both at once. If the listener feels the wrongness a half-second before the safety arrives, you've already lost them — they've recoiled. If they feel only the safety and the wrongness never registers, there's nothing to laugh at. McGraw and Warren are strict about this in the 2010 paper: humor occurs when and only when all three conditions hold together. Miss the timing on the overlap and even a well-built joke just won't fire.
So gather those three for a second, because they do all the work from here. Wrong but harmless. Both at once. Too safe and it's boring, too hostile and it's offensive, out of sync and it's nothing.
Now to the dinner-table puzzle. Why does the same Hummer joke split the room? Because "benign" isn't a property of the joke. It's a property of the listener. The wrongness might be fixed, but whether it feels safe depends entirely on who's hearing it — and McGraw's lab has mapped exactly which dials move it.
One dial is commitment to the norm being broken. McGraw and Warren found that the people who laughed easiest at the church-and-Hummer setup were the ones who weren't religious. They had no deep stake in the sacred norm being poked, so the poke felt benign. To them: harmless and a little absurd. To a devout churchgoer, the same words press on something load-bearing, and the violation stays malign. The lab found the same pattern with sexist jokes — in their research, men tended to find them funnier than women did, and the theory's explanation is uncomfortable but clean: if a norm's violation costs you nothing, it's easier to read as safe. That's not a verdict on whether the joke is good. It's a map of why your audience is split.
The second dial is the one with the famous nickname, and it's the most useful lever you've got. Psychological distance. You already know the folk version — comedy is tragedy plus time. McGraw didn't just nod at that saying; his lab ran it as an experiment. In a 2014 paper in Social Psychological and Personality Science, he and his colleagues studied what they bluntly titled "the rise and fall of humor," tracking how funny people found tragedies as time passed. And the findings were sharper than the cliché. Distance isn't only about time. A violation feels more benign when it happens to someone else, or far away, or long ago, or in a way that doesn't quite seem real.
Here's the twist that makes this genuinely useful, and it's the title of another McGraw paper — "too close for comfort, or too far to care." Distance isn't a one-way slider where more is always better. Too little distance and a mishap is just a threat: a real injury happening to you, right now, is not funny, it's an emergency. But too much distance and the violation evaporates entirely — it's so remote, so abstract, so long ago that there's no wrongness left to resolve, and you get a shrug instead of a laugh. The sweet spot is a window in the middle. Close enough that the violation still registers, far enough that it feels safe. A stubbed toe is funniest to the person who didn't stub it. A disaster from a century ago can be a punchline; the same disaster this morning is just grief.
That's the part most people get backward. They assume the goal is to make a joke as safe as possible. But maximum safety kills the joke just as dead as maximum offense — you've sanded off the violation entirely. The craft isn't minimizing the edge. It's calibrating it.
And calibration is where this becomes a skill you can actually run, instead of a theory you can admire. Think of every remark you're about to make as a dial with two knobs. One knob sets how big the violation is — how wrong, how pointed, how close to a nerve. The other sets how much you're signaling that it's safe — your tone, your smile, the obvious affection underneath it, the fact that you're aiming at yourself instead of the person across from you. A good comic isn't pushing one knob to the max. They're reading the room and turning both until the two readings overlap in that narrow funny window.
Self-deprecation is the cleanest example of calibration in action, and it's worth seeing why it works mechanically. When you make the violation about your own flaws, you've solved the benign problem before you even open your mouth. Nobody in the room feels threatened by you mocking yourself — there's no one to defend, no one getting punched down at. The wrongness is real, you're admitting something genuinely unflattering, but the safety is total. That's not just "nice" humor. It's the most reliable way to keep a violation benign when you don't yet know where your audience's lines are.
Which is the practical heart of the whole thing. The same exact line is benign to one person and malign to another, depending on what they're committed to, how close the wound is, how much they trust that you mean no harm. So before you push the edge with a stranger, you turn the violation knob down and the safety knob up — small jokes, gentle targets, plenty of warmth — and you watch. The more you learn about where their lines actually sit, the more precisely you can push. Reading those reactions in real time is its own skill, and a later part of this course is built entirely around it. For now, the move is simply this: treat edge as something you dial, not something you guess.
So if someone stopped you right here and asked why a remark that destroyed at one party died at the next — what would you say? … The words didn't change. The listener did. And benign lives in the listener, not in the line.
Strip away the lab names and the papers, and three things are doing the real work. A joke needs something genuinely a little wrong, or there's nothing to laugh at. It needs that wrongness to feel safe at the very same instant, or you've just upset someone. And whether it feels safe isn't fixed — it slides with how close the wound is, how much the listener cares about the norm you poked, and how much they trust you. Comedy is tragedy plus time, yes. But it's also tragedy plus distance, plus the right audience, plus a tone that says we're okay here.
That reframe is what turns "some jokes just don't land" into something you can fix on purpose. The joke that bombed wasn't badly written. The violation was probably fine. What was off was the overlap — too much edge for that listener, or not enough trust built yet, or a wound still too fresh to laugh at. And if benign and violation have to hit at exactly the same instant, then the thing that decides whether they overlap isn't just what you say. It's when you say it — which is where this gets stranger, because it turns out timing might matter even more than the words themselves.