January 20, 2018
If you pick up a book on sociology you'll find some mention of the concept of laughter and how it is an important part of our social lives, but you'll find very little written about the evolutionary biological purpose of laughter, or any kind of approach toward providing a more scientifically quantifiable definition of what is or is not humorous.
Before I go into more verbose detail, I'll cut to the chase and propose a specific and somewhat quantifiable definition of humour: The purpose of laughter, and humour in general, is to communicate the presence of a significant, but non-threatening inconsistent mental or emotional state in one or more individuals. The inconsistency could be between oneself and another individual, or between two or more individuals other than oneself. In other words, it is the most primitive method of communicating to someone else any or all of the following: "You/they are not thinking what you/they should be thinking." "Your/their prediction was wrong." "You/they are not smart." "You/they did something I didn't expect." "Your/their emotional response isn't what it should be".
Now, you'll probably tell me that this definition is 'obvious' and that it's a waste of time to write about something that is 'so obvious', but I think that if you want to really understand what's going on in other people's heads, then you need to be more skeptical of 'obvious' explanations.
If you do a Google search for something like "Why do babies laugh?", or "What is the purpose of humour?" you'll find answers like "Crying is a signal they want something to change, while smiling or laughter is the opposite - it says keep doing what you're doing", or "The purpose of humor is to laugh at ourselves.". Sigmund Freud's explanation for laughter was that its purpose was to 'relieves pent-up nervous energy'[1]. Charles Darwin sees it as a form or re-balancing or dissipation of pent up emotions: "The excess must charge itself in some other direction."[2] In my opinion, these are terrible answers that don't make any progress at explaining the true common denominator of all forms of humour or laughter.
Any meaningful description of the purpose of humour or laughter should be capable of explaining the following situations:
With respect to the above situations, I'll make an attempt at connecting them to general definition I've stated in the introduction:
I don't have much experience with babies myself, but from reading anecdotal descriptions of what makes babies laugh, and the descriptions written by Charles Darwin[3], there seems to be several common causes of baby laughter. The two most common are tickling and the Peekaboo game. Interestingly, laughter from tickling and the peekaboo game seem to develop at different times, which likely provides information about the internal development process of a baby's perceptual understanding of the world. Tickling, as a source of humour, is one of the things that I've found the most difficult to explain as a source of humour, but I'll claim that its origin is as follows: Ordinarily, the intense sensation of being touched in sensitive areas causes illicits a defensive instinct. When these areas are stimulated by someone who is trusted (mother/father), the natural instinct to be defensive of the touch conflicts with the natural instinct of having feelings of affection toward the loved one. This creates an inconsistency that we vocalize in the form of laughter.
In a similar vein, I would reason that the cause of laughter from the Peekaboo game occurs for one of several reasons: When a baby is just beginning to learn the concept of Object Permanence, the rather surprising idea that someone or something could 'appear out of nowhere' is surprising, but simultaneously non-threatening when it comes from a loved one. Gradually, as the baby's model of object permanence develops further, the inconsistent (and therefore funny) aspect of peekaboo would change from being surprised about the unexplained appearance and disappearance of objects, to the error in making predictions about exactly when the next appearance will occur and the pseudo predatory role play of hiding and surprising.
In accordance with the definition provided earlier, I claim that in this case, the laughter is a social signal that an individual sends out to make sure that others understand that they aren't serious. Obviously, if you make a joke about doing something potentially illegal or dangerous, it can be important to confirm to others that you're not serious, but this extra confirmation is only necessary if others don't signal that they understand your joke by laughing themselves.
The analysis in the last paragraph is something fairly trivial that most adults figure out in some way or another, but another related question would be "Why are there people who consistently make jokes that aren't funny, then laugh at them?" and a related question: "why are these people seen as immature or annoying?". Drawing on my definition, I believe I can explain these two questions as well: An individual who is unsure of the thoughts, feelings, and opinions of others can 'probe' how others process a hypothetical idea by consistently making jokes and judging the response of others. If the response is laughter, others process the information in the same way. If there is no laughter, their thought process is different from their own, and it becomes necessary to signal their experiment by laughing.
Unfunny jokes that occur too frequently can therefore represent an individual with an under-developed, mal-developed, or simply has an incompatible mind that does not have the same perception or thought process to ones own. In other words, someone who is "not on the same page".
This one is fairly easy to answer given the above definition: Given that people who are mentally ill have abnormal perception or processing of the world around them (by the definition of mental illness), this would lead a mentally ill person to frequently have thoughts that were inconsistent with those of the majority of the population, therefore triggering humour for them. In addition they would also be more likely to have incorrect intuitions about the thoughts and feelings of others, which would lead them to perceive non-existent inconsistencies between the perceived thoughts of others.
In keeping consistent with the definition, laughter is the social signal that lets others know you're aware of a mental or emotional inconsistency. If you're aware that someone else is making a joke, but you find their proposed inconsistency trivial or relatively insignificant, you would normally not laugh. However, you know that the joker has an expectation of laughter from you, and an absence of laughter could be misconstrued as a confirmation by you that the proposed ideas in the joke are normal. Therefore, feigned laugher is necessary to confirm to the joker that you don't in fact agree with their joke ideas, while keeping them under the perception that you're both "on the same page".
I think that the explanation for this section would be similar the same question proposed about those who are mentally ill, since it isn't too controversial to suggest that the effect of various drugs is often similar to mental illness or developmental disabilities. However, I think that simply suggesting that it's identical to the explanation for mental illness would be selling this topic short of a good research opportunity: If my definition has any merit, it should be possible to identify drug specific differences in the perceptual and mental experiences that trigger laughter on those who are under the influence of drugs. For example, drugs like alcohol suppress the central nervous system, but drugs like amphetamines do the opposite. Given these differences, it seems reasonable that one could come up with hypotheses on what alcoholics would find funny, versus amphetamine users based on the scienficially known differences in how these drugs alter perceptual experience. I haven't researched the topic, but it would be interesting if someone else did!
Nervous laughter is usually not observed when someone has been fearful for a prolonged period of time, but it is usually seen in situations with a rapid onset or change in the level of stress or fear. In keeping with the definition of laughter as a form of communication, I would claim that nervous laughter is effectively used as a 'bluff' to communicate to others that you're aware of the inconsistency of your situation (things are dangerous/uncertain/stressful when they should be easy and relaxed), but that that you perceive them as non-threatening. This would effectively act as a protection mechanism that socially signals, "Yeah, this situation is a bit dangerous, but it's whatever." when in fact you're incredibly scared. If your bluff works, there would be clear benefits in a confrontation, or in moderating the fear response of your peers.
According to Wikipedia, some animals such as apes, rats and dolphins do have vocalizations that are seem similar to laughing and they occur in situations that resemble tickling. Interestingly, there isn't a lot more written than this about laughter in animals. Perhaps, this is indicative of the more highly evolved capacity of humans for complex social communication: We can communicate complex mental states through the simple signal of laughter or no laughter, and maybe this is one of the major characteristics that separates us from animals.
The literary concept of 'dramatic irony' is very consistent with the definition proposed above: As defined in the Oxford English Dictionary, dramatic irony is "a situation in a play when a character’s words carry an extra meaning to the audience because they know more than the character, especially about what is going to happen".
Given the social importance of humour it is quite surprising how little humour has been studied. In general, if a member of your social group never 'gets' all the popular jokes that will usually make it impossible for them to gain high social status. Conversely, being on the butt end of a joke that you don't understand is probably one of the worst social feelings. It is no surprise that so much of our culture is filled with endless jokes about everything from trivial daily activities to long complicated situations: We crave the confirmation that we can understand the mental states of those around us, and seek to ensure that our thought process is compatible with those around us. If it wasn't, we would be isolated and unable to communicate with the rest of social society.
There is a YouTube video where Jerry Seinfield describes his thought process behind how he writes jokes. One of his closing thoughts is that "In my world, the wronger something feels, the righter it is. To waste this much time on something this stupid, is, that felt good to me.".
In the context of computer programming, the concept of a 'diff' is a simple summary of the differences between two of anything. It could be two examples of computer code, two traces of a program state, or two examples of the output of a program. In general, I have found the primitive concept of diffing to be a common denominator to isolating and solving any kind of problem. If you can find a way to detect and express the differences between the program that you want and the program you have, you can build or fix anything. In the same way, I think of the concept of humour as humanity's most primitive diffing algorithm for social situations. It is an extremely fundamental tool that we use to establish what is right and wrong, what is expected or unexpected, and what is smart or stupid.
Very soon, if not already, we will start asking questions related to humour and artificial intelligence. Can an AI figure out how to say things that are funny? I would say: "Of course!".
In fact, I think (and would propose to fellow computer scientists out there) that if you want to build an AI that can understand social context, using the concept of humour as a primitive 'diffing algorithm' would probably be an excellent heuristic. If you can get an AI to figure out when exactly something is funny and quantify the inconsistency as an RMS error, I would say you're on you way to having a great algorithm that can update the weights that regulate social behaviour!
Some people probably think that making good jokes is something AI will have difficulty with, but I don't think it will be that hard and we'll have to start dealing with it sooner than we think. What I also think will happen, is that AI will easily be capable of revealing to us, through its sense of humor, inconsistencies in our own thoughts and behaviour that we don't want to think about. If you want me to be more specific, there are a number of contentious political and social ideas out there that are ripe for humour because of how obviously ridiculous they are. The trouble is, it's hard to navigate them without invoking riotous laugher in one half of the population and vehement anger in the other half. In these cases, the sense of humour of our new AI will need to be 'patched' to re-describe our own inconsistencies as though they were in fact consistencies. Only when an AI has been programmed to be as inconsistent as a real human being will it be truly worthy of calling 'human'.
[1] https://plato.stanford.edu/entries/humor/
[2] http://darwin-online.org.uk/converted/pdf/1897_Expression_F1152.pdf Page 198 THE EXPRESSION OF THE EMOTIONS IN MAN AND ANIMALS, Charles Darwin.
[3] http://darwin-online.org.uk/converted/pdf/1897_Expression_F1152.pdf Page 199 THE EXPRESSION OF THE EMOTIONS IN MAN AND ANIMALS, Charles Darwin.