Posts
Comments
I replied breifly before but I can't find it now, maybe cause I'm new. Also maybe cause it's 2am and I've just woken from a coffee bender induced nap. I feel I'm doing the conversation a disservice by writing on the fly, at 2am on my phone, but I feel compelled to say that I think humour (and laughter) are vital to the development and regulation of agi.
I love all of your reasoning regarding the causes of Laughter. Every argument seems sound to me and every counter argument I just read is welll answered already in your theory or by the person replying. In your search for the physiological mechanisms, have you considered that a particular region of the brain may not be the location? Laughter seems intrinsically linked to learning, and hence to the most basic neural proceses. Perhaps the mechanism is more a diatributed signalling process than any particular single structure.
One thing to consider about your model of humour is that all of the ingredients and nuances suggested by yourself and others for a and b, along with possibly c, are things that are the "sudden knowledge" in the quote I love about what laughter is.(please someone remeber where that is from for me!).
When our brains encounter knowledge, one expects the learning system of the brain to operate in a way that stimilates the persistence of that knowledge.The usual suspect here is the molecular signalling and feedback of the reward system. But what if that system becomes overloaded?
The startle/orienting process you describe, in all its forms, is knowledge being collected. The importance/significance is increased when danger is present, or when strongly held notions are upended. This significance should result in a stronger learning response than usual, and if the timing is right and the signal is strong enough, then a messaging or control molecule cascade or overabundance may ensue as inhibiton/depletion of the molecular signalling fails to keep up with production.
Tickling might not necessarily seem to be part of a learning/reward system. But the brain, especially oarts developed early in human(or mammalian) evolution, won't know that. To the brain, neural activity (novel neural activity in particular) is treated like learning, whether that activity is from sensory nerves stimulated physicaly, or brain neurons stimulated by thoughts and ideas.
Physical sensory nerves in ticklish areas might also provide a novel neuronal stimulation as the information from these sensory nerves overload the learning reward system. Ticklish areas, as well as being vulnerable (perhaps because they are vulneable) are possibly less frequently active and more abundant. Or perhaps the nature of the stimulation (tickling often needs to be done lightly to be effective) means these nerves can remain active longer/repeatedly since their potential isn't discharged or inhibited completely. Either way, if these nerves can provide more abundant, rapid or novel signalling than usual sensory input to the brain, they qualify as sudden knowledge.
Anything that can be considered knowledge of an abrupt and definitive nature, which produces a signal above a certain threshold, wil make us laugh. This threshold is highly individual, and changes as our perception of what is novel changes, as our worldview alters what startles us/how we orient, and even if we have drugged our poor brains with weed or alcahol.
Beauty works in a similar way, but Instead of a sudden and overwhelming stimulus being resolved to specific knowledge, it is the knowledge itself which is overwhelming. The resolution of a stumulus with a deep or strongly held knowledge produces the cascade here, causing an overwhelming reinforcement of held knowledge rather than new. But thats another two am bunch of dribble so I'll not say any more of how beauty works.
Understanding why and how laughter works (and beauty also) may be a key concept in getting an agi to develop curiosity. If an agi can be developed to be curious for curiosities sake, then that agi becomes not just a machine that can learn,, but one that becomes truly intelligent.
More importantly, concepts such as beauty and humour take an agi towards developing empathy and compassion. Despite the complexities and inconsitencies such 'emotional' traits bring, I think they are imperative for agi safety. So I think your musings ar far more than a distraction, and quite an interesting and important area to investigate.