Has anybody written about humor as a sign of a successful encryption strategy?
I think that a good joke might be a lot like a sweet encryption, and I’d like to explore this notion:
The sweet spot of jokes is actually the same as the sweet spot of sweet cryptography schemes.
Consider two failed joke forms:
- The non sequitur: “A priest, two rabbis, and a Zen monk walk into a bar. Punchline: None of them eat rutabagas.”
- The supra sequitur: “What’s a lot like a horse and has black and white stripes?Punchline: A zebra.”
Aside: I actually think these are both quite funny, but they are funny as metajokes: they violate the constraints of joke-dom rather neatly. They force the listener to circumambulate the Gricean Maximal Limits before returning to confront the teller with a withering, sidelong glare: the listener has finally settled on the Grice Hypothesis of Last Resort (“Uncooperative Communicant”).
These two jokes fail to match these premises:
If the punchline is predictable, it’s a bad joke. The punchline must be difficult to predict.
If the punchline doesn’t “make” the joke quickly, it’s also a bad joke. Fitting the punchline to the setup must be obvious in retrospect.
Let’s spell out the analogy:
message:joke :: encrypted form:setup :: decryption key:punchline
By this analogy, we can say the same things about encryption strategies:
If the decryption key is predictable, it’s a bad encryption. The decryption key must be difficult to predict.
If the decryption key doesn’t “make” the message interpretable quickly, it’s also a bad encryption. Fitting the decryption key to the encrypted message must be obvious in retrospect.
Humor seems to live in the “second law one-way” sweet spot that good cryptographic approaches use: it’s easy to confirm correctness of a given solution, but difficult to predict a correct answer.
Does anybody know citations pointing to people publishing in this direction? Either thinking about humor as cryptographic success, or thinking about crypto as “setup + punchline”?