You’ve just enjoyed a standup comedy routine or watched a comedy in the theater. Maybe, you even laughed until you cried. You feel good.
Why do humans laugh? What causes the laughter? What is the state of laughing science?
E.B. and Katherine White observed that:
“Humor can be dissected, as a frog can, but the thing dies in the process and the innards are discouraging to any but the purely scientific mind.”
You have break some eggs to make omelets; can we sacrifice frogs for humor?
A professor at the University of Alberta’s Department of Psychology, Chris Westbury, immodestly claims that he has developed a model of humor by using artificial intelligence:
“My model will tell you in advance what things people might find funny…Nobody has really done that before on a large scale.”
Westbury and colleagues added 45,000 English words to an existing study of 5000 words. All of the words were fed into an algorithm to predict each word’s potential for humor. They “discovered” two stunners: a word is funny based on its meaning and sound.
“Our models are complicated. They have a lot of predictors in them,” Westbury said. “It’s a first step toward quantifying humor, and that’s important because it shows that none of the theories we have [for humor] are good enough.”
Thanks to artificial intelligence we can do humor by the numbers. We can create an algorithmic formula so that anthropomorphic machines (aka robots) can toss off the quick quip, the sly witticism that makes us feel like we’re with a friend. A human friend.
Westbury also identified what he calls the ten funniest words in the English language.
Is one word inherently funny? Can one word make you laugh?
Are you laughing now? You should be in stitches according to the algorithm’s humor predictions. For you have just encountered the algorithm’s ten funniest words in the English language, each one appearing in cameo between paragraphs—entr’acte solos.
I’ll bet your jiggly puffball that you and your bubby barely cracked a wriggly giggle through that list.
Why? Because there is no standard of humor to rely on. Humor is not static, it is not composed of algorithmic building blocks. What is funny today is not funny tomorrow. In the early 1900s, someone who was funny were described as a real “caution.” In the 1950s, the funny guy would be called a real “panic.”
We can’t predict humor no matter how many thousands of words we feed into an algorithm. Predicting humor kills the frog before its first guffaw: when you know the joke, or you can see the punch line coming from a mile away, it’s not funny. What makes you laugh is the unpredictability, the twist in the narrative that you didn’t see coming.
Nevertheless, Westbury and crew should be congratulated for exploring one of the great mysteries of the human mind: why do we laugh? If we can find a pattern to what makes us laugh, perhaps we will understand why.
But, running 50,000 words through an algorithm is small potatoes. Bob Hope had 528,000 jokes filed in twelve four-drawer cabinets stored in a fire-proof Mosler bank vault. Hope labeled five of the cabinets “keepers” which he donated to the Smithsonian.
Hope ran a joke factory, spending millions of dollars on over 100 different writers. He was one of the most popular and financially successful comedians of the 20th century. You could say he was a human algorithm of jokes.
The jokes and gags must have been hilarious during his hey-day. You can read them on-line.
But now, taken out of the context of his era, his persona, and his delivery, the humor has gone stale.
An algorithm could be devised to delve into Hope’s file cabinets to determine his humor patterns. (Even with 100 different writers, you can still recognize a Bob Hope style joke.)
We could postulate what made Hope funny to the people of his time.
We might extract some understanding of the frog’s past wriggles. But, that frog died a long time ago. And we have no idea what new frogs will hop into our lives. Predictable humor is dead humor.
Algorithms or not, we still can’t predict humor. And that’s the way it should be.