The most advanced form of humour in the world.
You may or may not believe me, but they do take a considerable amount of work and preparation to perform.
I have a few mental models/analogies that seem to help with pun construction and execution.
The simplest of the three has to do with how people construct sentences from concepts. My experience is that most people only pay attention to the concepts that their brains create, and don’t watch for the words that are emitted. You can test this by asking someone to repeat what they said, and watching how often they use the exact same words.
*-----------* *------------* | Concept | | Concept | | |---->| to word |---->Speech | generator | ^ | translator | ^ *-----------* | *------------* | | | Where people | | pay attention--/ | | Where punsters | pay attention---------------------/
The second model is that of ‘hash functions’. Even though the brain is a neural net, many of its attributes can be modeled as hash functions. In this context, a ‘hash function’ is a very quick retrieval of something (an idea, concept, memory), based on a trigger. But this often only works in one direction. If you ask someone what they were ate yesterday for breakfast, they may not be able to tell you. But if you ask them the last time they ate toast, they would be far more likely to say ‘yesterday, at breakfast’. This can cause issues of the ‘why don’t you remember that?’ and ‘I already told you that’ sort when people have different hash functions, and associate things differently.
The third model (and my favourite!) is a poor description of part of how I listen for puns. I’ve talked about word and sentence rotation recently, but much of that is the slower, ‘software’ way of manipulating sentences to extract verbal humour. This next analogy seems to be much more hardwired.
The analogy is of a record player playing a record. The record player is my ear and word processing apparatus, the record is the incoming vocal stream. So far, so good. What it feels like I do is to lift the needle slightly off the record. I then engage my sound->word prediction, and come up with as many words as I can that sound similar to those that are being and will be used. So, I’m turning my exact word matching into fuzzy or inexact matching, then using the results to construct puns.
Other models of how to pun? Let me know in the comments below!
Remember how you need to be able to predict the exact words someone will use.