WEBVTT - Ep74 "Why do we laugh?" 0:00:05.160 --> 0:00:08.719 From the brain's point of view, What is humor when 0:00:08.720 --> 0:00:12.719 something is funny? Why do we breathe in and out rapidly? 0:00:13.320 --> 0:00:17.720 Do other animals laugh? Why do most jokes come in threes? 0:00:18.440 --> 0:00:22.599 What do mystery novelists, magicians and comedians have in common? 0:00:23.079 --> 0:00:30.200 And could AI be truly funny? Welcome to Inner Cosmos 0:00:30.200 --> 0:00:33.280 with me and David Eagleman. I'm a neuroscientist and author 0:00:33.320 --> 0:00:36.880 at Stanford and in these episodes we sail deeply into 0:00:36.880 --> 0:00:40.760 our three pound universe to understand why and how our 0:00:40.800 --> 0:00:45.320 lives look the way they do. Today's episode is about humor. 0:00:52.920 --> 0:00:55.880 Although humor is one of the most valued aspects of 0:00:55.920 --> 0:00:58.880 our lives, you've probably heard very little about it from 0:00:58.960 --> 0:01:02.720 the scientific point of view, and that changes today. So 0:01:02.840 --> 0:01:07.120 why do we laugh? After all? It's very strange behavior. 0:01:07.680 --> 0:01:10.600 Imagine that you are a space alien and you are 0:01:10.640 --> 0:01:14.800 studying humans. So through years of study, you note that 0:01:14.920 --> 0:01:18.200 humans seem to have a relatively high bandwidth channel of 0:01:18.280 --> 0:01:23.360 communication by an extraordinarily rapid and precise use of their 0:01:23.920 --> 0:01:27.160 larynx and their vocal cords. They move their tongue and 0:01:27.200 --> 0:01:29.840 their lips very rapidly. They blow air at just the 0:01:29.920 --> 0:01:33.399 right moments, and in this way they have language. They 0:01:33.440 --> 0:01:37.160 communicate to each other at about thirty nine bits per second, 0:01:37.160 --> 0:01:40.400 which is a pretty good transmission rate and way higher 0:01:40.440 --> 0:01:43.120 than any other animal species as far as you the 0:01:43.200 --> 0:01:46.640 alien can tell. But there's something else. These human creatures 0:01:46.680 --> 0:01:49.880 do something that's very strange. Every once in a while, 0:01:50.440 --> 0:01:54.720 they do a much lower frequency, lower bandwidth thing where 0:01:54.760 --> 0:02:00.240 they make a series of rhythmic, involuntary contractions of their diaphragm, 0:02:00.640 --> 0:02:06.639 which causes a sequence of vocalized exhalations, usually a series 0:02:06.720 --> 0:02:11.880 of short repeated vowel sounds like ha or he or ho. Now, 0:02:12.040 --> 0:02:16.400 you notice that this typically begins with some language or 0:02:16.440 --> 0:02:20.760 event that triggers the brain's reward centers and lights up 0:02:20.760 --> 0:02:23.880 something about the emotional networks, and that triggers the activation 0:02:24.040 --> 0:02:28.920 of facial muscles, particularly around the mouth and eyes, And 0:02:29.040 --> 0:02:32.280 often the human's mouth will open up wide and the 0:02:32.320 --> 0:02:35.240 corners of the lips will get drawn back and upward, 0:02:35.840 --> 0:02:40.160 and the respiratory starts doing these rapid cycles of inhalation 0:02:40.280 --> 0:02:44.840 and exhalation, which throws a wrench in the normal breathing patterns, 0:02:45.360 --> 0:02:49.000 and the heart rate increases and the blood pressure rises, 0:02:49.680 --> 0:02:53.120 and the eyes might start to produce tears, and the 0:02:53.160 --> 0:02:56.600 skin around the face might start turning red because of 0:02:56.639 --> 0:03:01.640 the increased blood flow. Sometimes the whole body shakes or quivers, 0:03:01.720 --> 0:03:06.720 especially in the shoulders and torso, and this physical manifestation 0:03:06.800 --> 0:03:10.760 that can spread to the limbs, leading to movements like 0:03:11.120 --> 0:03:14.600 clapping or slapping or bending over. Now, the question you 0:03:14.639 --> 0:03:18.360 ask as the alien is why do they do this 0:03:18.520 --> 0:03:22.560 strange laughing thing? Is there an evolutionary reason behind it? 0:03:22.720 --> 0:03:26.480 Do any other animals on the planet do it? So 0:03:26.639 --> 0:03:30.840 you go back to your alien university to present your findings, 0:03:31.200 --> 0:03:33.960 and the faculty there drills you with questions that you're 0:03:33.960 --> 0:03:37.240 supposed to know the answer to, like why do humans 0:03:37.400 --> 0:03:41.480 go to clubs and movies with the hopes that they 0:03:41.520 --> 0:03:45.640 will trigger a lot of these diaphragm contractions And if 0:03:45.720 --> 0:03:49.800 they get fewer contractions than expected, do they seem disappointed? 0:03:50.480 --> 0:03:53.360 Why do they get together with other humans that they 0:03:53.440 --> 0:03:56.600 trust to stick food in their mouths and hope to 0:03:56.680 --> 0:04:00.720 do these strange diaphragm contractions while eating, even though it's 0:04:00.840 --> 0:04:03.440 dangerous and could cause them to choke on their food. 0:04:03.680 --> 0:04:06.400 So we're going to take this in a few acts. First, 0:04:06.400 --> 0:04:08.960 we're going to address what is humor, And in the 0:04:09.000 --> 0:04:12.760 second act we'll ask why do we laugh? In other words, 0:04:12.760 --> 0:04:14.960 why do we do it out loud instead of silently. 0:04:15.160 --> 0:04:18.600 So let's start with the general notion of comedy about 0:04:18.640 --> 0:04:21.880 something being funny. As it turns out, people have been 0:04:21.960 --> 0:04:25.159 wondering about this for a long time, and the first 0:04:25.200 --> 0:04:30.000 known theories about humor go back to Plato. Now, Plato 0:04:30.440 --> 0:04:34.120 had a very dim view of laughter. He generally saw 0:04:34.160 --> 0:04:37.880 it as something malicious that people do. In his view, 0:04:38.080 --> 0:04:42.760 audiences laugh at people who have ignorance about themselves, who 0:04:42.800 --> 0:04:46.200 imagine themselves to be richer than they are, or better 0:04:46.240 --> 0:04:48.719 looking than they are, or more virtuous than they are, 0:04:49.360 --> 0:04:51.880 and we laugh at them, which he found to be 0:04:52.000 --> 0:04:56.719 a morally objectionable thing to do. Now, his student Aristotle 0:04:56.880 --> 0:05:00.800 appreciated wit in conversation, but he agreed with Plato that 0:05:01.000 --> 0:05:05.000 people laugh at the misfortunes or mistakes of others, which 0:05:05.400 --> 0:05:10.320 reinforces their own sense of self worth and status. Think 0:05:10.400 --> 0:05:13.400 of an audience laughing at a character in a comedy 0:05:13.760 --> 0:05:17.279 who's consistently making foolish mistakes. So this is what's known 0:05:17.320 --> 0:05:22.320 nowadays as superiority theory, and the idea is that comedy 0:05:22.360 --> 0:05:26.480 stems from feelings of being better than other people, and 0:05:26.520 --> 0:05:29.520 no one really challenged this idea that that's what comedy 0:05:29.560 --> 0:05:34.040 was about. Until the mid eighteenth century people started to think, 0:05:34.600 --> 0:05:37.719 wait a minute, you don't really need feelings of superiority 0:05:37.760 --> 0:05:41.359 to laugh at something. Just think of a funny poem 0:05:41.440 --> 0:05:46.160 or pun. And on the flip side, we often feel superior, 0:05:46.240 --> 0:05:48.240 like the way you might feel superior to a horse 0:05:48.320 --> 0:05:51.560 or a squirrel, but that doesn't mean we laugh at them. 0:05:51.760 --> 0:05:55.200 So other theories began to blossom, and now there are 0:05:55.640 --> 0:05:59.360 many theories of humor. One of them, for example, is 0:05:59.560 --> 0:06:04.000 relae theory. This was originally proposed by A Lord Shaftsbury 0:06:04.040 --> 0:06:08.279 in the seventeen fifties, and he suggested the slightly strange 0:06:08.320 --> 0:06:12.520 idea that we laugh to release animal spirits that have 0:06:12.600 --> 0:06:15.599 built up inside us. Now, this doesn't make much sense 0:06:15.640 --> 0:06:19.320 in our modern vocabulary, but later proponents of this idea, 0:06:19.480 --> 0:06:23.400 like Sigmund Freud and Herbert Spencer, they cast this same 0:06:23.480 --> 0:06:27.159 idea in a different light. They said, look, humor is 0:06:27.200 --> 0:06:32.919 a way to release psychological tension or suppressed emotions, like 0:06:33.000 --> 0:06:35.599 think of laughing at a joke that touches on a 0:06:35.600 --> 0:06:40.200 taboo subject. So laughter serves as a release valve for 0:06:40.400 --> 0:06:43.960 pent up energy in your nervous system, and in this view, 0:06:44.000 --> 0:06:47.279 it's a purely physiological issue. And you see this happen 0:06:47.360 --> 0:06:51.720 sometimes when someone laughs at a totally inappropriate moment, like 0:06:52.040 --> 0:06:55.440 someone has died, and the excess energy comes out in 0:06:55.440 --> 0:07:00.599 a way that looks different from what they're feeling internally. Relatedly, 0:07:00.680 --> 0:07:04.440 by the way, Freud also suggested that humor and laughter 0:07:04.560 --> 0:07:07.880 might act as a defense mechanism, and in this way 0:07:07.920 --> 0:07:12.920 it allows people to deal with uncomfortable emotions or social tensions, 0:07:13.320 --> 0:07:15.800 and it can reduce stress and anxiety. But there are 0:07:15.800 --> 0:07:18.920 other ways that people look at humor. One very common 0:07:18.920 --> 0:07:21.920 school of thought looks at humor as a way of 0:07:22.080 --> 0:07:27.240 indicating play. It says, look, play is absolutely essential for 0:07:27.520 --> 0:07:29.920 learning how to do real things in the world. And 0:07:29.960 --> 0:07:34.800 so the idea is that humor engages all this cognitive 0:07:34.840 --> 0:07:40.160 practice and social practice in a playful context. For example, 0:07:40.160 --> 0:07:44.000 look at the way that children engage in silly wordplay 0:07:44.400 --> 0:07:49.960 or crazy imaginative scenarios. That stuff cracks them up. They're playing. 0:07:50.760 --> 0:07:52.920 And there's a closely related way of looking at this too. 0:07:53.040 --> 0:07:57.120 When that evolutionary psychologists like the idea that humor and 0:07:57.200 --> 0:08:01.440 laughter is all about social bonding. Humor is a tool 0:08:01.640 --> 0:08:06.240 for strengthening the bonds within a group. You build cohesion 0:08:06.280 --> 0:08:09.280 this way, you build a sense of community. So just 0:08:09.280 --> 0:08:12.520 think about inside jokes that you do with your friends 0:08:13.200 --> 0:08:17.960 that reinforce your group identity and your camaraderie. And I 0:08:18.000 --> 0:08:21.960 think that the slang that each generation introduces is a 0:08:22.000 --> 0:08:24.880 form of the same thing, making sure that the adults 0:08:24.880 --> 0:08:27.320 are on the outside and you and your pals are 0:08:27.320 --> 0:08:31.080 on the inside. It's a way of bonding. Now, the 0:08:31.200 --> 0:08:35.400 ideas about humor don't end there. Other thinkers like Marvin 0:08:35.440 --> 0:08:40.280 Minsky and Jerry Saul's focus on a different aspect of humor, 0:08:40.320 --> 0:08:45.640 the cognitive aspect. They care about the mental processes involved 0:08:45.640 --> 0:08:49.160 in understanding and appreciating comedy. So they're looking at the 0:08:49.679 --> 0:08:54.120 cognitive effort required to understand some cognitive play on words. 0:08:54.280 --> 0:08:57.280 So take a joke like this today. I found a 0:08:57.320 --> 0:09:01.560 book with the title how to Solve fifty of Life's Problems, 0:09:01.960 --> 0:09:04.520 So I bought two of them. So the last sentence 0:09:04.559 --> 0:09:07.680 comes as a surprise, and the brain goes back and 0:09:07.720 --> 0:09:11.720 tries to reinterpret what it heard. There's a puzzle in there. 0:09:12.240 --> 0:09:17.640 Which forces you to reshape your interpretation, or specifically, your 0:09:17.760 --> 0:09:22.400 interpretation of the speaker's interpretation. And then it falls into 0:09:22.400 --> 0:09:24.840 place that there's another way to see this, and it 0:09:24.960 --> 0:09:29.520 strikes us as funny because it's both nonsensical and a 0:09:29.559 --> 0:09:33.360 little bit sensical. And some other theories zoom in even 0:09:33.440 --> 0:09:35.440 more on the types of humor that have to do 0:09:35.520 --> 0:09:39.040 with linguistics, in other words, about the structure and use 0:09:39.080 --> 0:09:43.480 of language, like wordplay or jokes that rely on double 0:09:43.600 --> 0:09:48.040 meanings like I'm addicted to break fluid, but it's okay 0:09:48.120 --> 0:09:52.080 because I can stop at any time, or jokes that 0:09:52.160 --> 0:09:56.120 rely on homophones, like a photon checks into a hotel, 0:09:56.480 --> 0:09:59.080 the front desk asks if it has any luggage and 0:09:59.120 --> 0:10:03.000 it says, no, Oh, I'm traveling light. So the punchline 0:10:03.040 --> 0:10:05.800 is a phrase that can be reinterpreted and has two 0:10:05.880 --> 0:10:11.200 meanings that surprisingly fit. Now, fundamentally, your brain is always 0:10:11.200 --> 0:10:13.800 trying to squeeze the complexity of the world down to 0:10:13.880 --> 0:10:17.720 a single interpretation. And something about the back and forth 0:10:17.760 --> 0:10:22.520 here between two interpretations that can both work makes something funny. 0:10:22.920 --> 0:10:26.640 Or take a kid's joke like why aren't lobsters generous 0:10:27.000 --> 0:10:31.400 because they're shellfish? This relies on two sounds that are 0:10:31.440 --> 0:10:34.400 almost the same, so your brain entertains them both and 0:10:34.480 --> 0:10:38.960 finds humor in the surprise that they both work. So 0:10:39.440 --> 0:10:41.920 I've told you about two thousand years of people making 0:10:42.000 --> 0:10:46.160 theories about humor. It's about superiority, it's about releasing tension, 0:10:46.320 --> 0:10:50.120 it's about play, it's about social bonding or linguistics or 0:10:50.120 --> 0:10:53.040 so on. And the thing to appreciate here is that 0:10:53.080 --> 0:10:57.280 these aren't mutually exclusive. Each of these theories has something 0:10:57.320 --> 0:11:01.880 to say about some element of humor, provide a different lens. 0:11:02.520 --> 0:11:06.480 And what this proliferation of theories really highlights is the 0:11:06.600 --> 0:11:11.000 multifaceted nature of humor and all the processes cognitive and 0:11:11.120 --> 0:11:17.000 emotional and social involved in finding something funny. So if 0:11:17.040 --> 0:11:19.160 you are that space alien, you'd find there's not just 0:11:19.240 --> 0:11:22.360 one single answer for your colleagues who are peppering you 0:11:22.400 --> 0:11:26.600 with questions and trying to understand earth humor. Instead, you'd 0:11:26.600 --> 0:11:28.559 have to come to the conclusion that there are many 0:11:28.679 --> 0:11:32.839 angles to this strange behavior. But we haven't even gotten 0:11:32.880 --> 0:11:34.680 to the best angle yet. I want to zoom in 0:11:34.720 --> 0:11:37.680 on one theory in particular that I think sits right 0:11:37.800 --> 0:11:41.320 at the center of much of comedy, and that has 0:11:41.360 --> 0:11:46.800 to do with the violation of expectation. The engine of 0:11:46.960 --> 0:11:52.360 most jokes involves setting up an expectation and then breaking it. Now, 0:11:52.360 --> 0:11:55.200 why does that work. It's because the brain is constantly 0:11:55.600 --> 0:12:00.480 predicting what it thinks is likely to happen, and in 0:12:00.520 --> 0:12:05.000 certain circumstances, when the next steps seem clear, but then 0:12:05.000 --> 0:12:09.360 they're suddenly violated, that's often funny. So this theory of 0:12:09.440 --> 0:12:13.480 humor is called incongruity theory, and the school of thought 0:12:13.559 --> 0:12:17.440 is favored by philosophers like Emmanuel Kant and Arthur Schopenhauer, 0:12:17.920 --> 0:12:21.920 and it suggests that humor arises when there's a discrepancy 0:12:22.000 --> 0:12:24.520 between what you think is about to happen and what 0:12:24.640 --> 0:12:28.520 actually happens. So let's take an example. I saw a 0:12:28.559 --> 0:12:32.440 post on social media yesterday that read, my husband and 0:12:32.480 --> 0:12:35.160 I have reached the difficult decision that we do not 0:12:35.280 --> 0:12:39.280 want children. If anybody does, send me your contact details 0:12:39.280 --> 0:12:42.360 and we can drop them off tomorrow. So your brain 0:12:42.520 --> 0:12:46.400 starts off with a particular understanding and then that's tripped up. 0:12:47.080 --> 0:12:50.640 Or here's another example. My dog used to chase people 0:12:50.640 --> 0:12:53.040 on bikes a lot. It got so bad that I 0:12:53.120 --> 0:12:55.640 finally had to take his bike away. Now, jokes like 0:12:55.679 --> 0:12:59.640 this work in part because language is so low bandwidth. 0:12:59.760 --> 0:13:03.040 When you hear the sentence, you're making lots of assumptions 0:13:03.240 --> 0:13:08.360 unconsciously under the hood without ever realizing it. Another example 0:13:08.600 --> 0:13:11.800 is the way that we predict words ahead of hearing 0:13:11.840 --> 0:13:15.560 them again without even realizing it. So any comedy can 0:13:15.600 --> 0:13:18.560 take advantage of that. So I can tell you last 0:13:18.679 --> 0:13:22.559 month I got diagnosed as being colorblind. That really came 0:13:22.600 --> 0:13:25.520 out of the orange. Now there's only a surprise there 0:13:25.920 --> 0:13:29.600 because your brain was predicting what it thought was going 0:13:29.679 --> 0:13:32.720 to come next. And this is the reason so many 0:13:32.840 --> 0:13:36.960 jokes are structured in patterns of threes. The first guy 0:13:37.160 --> 0:13:41.320 does something, and then the second guy establishes the pattern. 0:13:41.840 --> 0:13:44.160 Then your brain has a clear prediction for what the 0:13:44.200 --> 0:13:47.000 third guy is going to do, and the comedy comes 0:13:47.040 --> 0:13:50.880 when he breaks that pattern. So a structure of three 0:13:51.520 --> 0:13:56.360 is the fastest possible path to setting an expectation and 0:13:56.400 --> 0:13:58.800 then swerving off that road. And by the way, this 0:13:59.000 --> 0:14:02.800 issue about the brain setting expectations, this is something that 0:14:02.840 --> 0:14:05.240 you can start seeing in the first few months of life, 0:14:05.720 --> 0:14:08.960 even well before infants have language. You can study this, 0:14:09.040 --> 0:14:12.560 and that's because when an infant is surprised by something, 0:14:12.559 --> 0:14:15.360 they'll pay more attention to it. So, for example, you 0:14:15.400 --> 0:14:18.440 put a teddy bear behind a screen, and then you 0:14:18.520 --> 0:14:21.640 remove the screen and now there's a rabbit there. Even 0:14:21.680 --> 0:14:24.440 a very young infant will stare and stare at that, 0:14:24.680 --> 0:14:27.920 indicating that even though they can't speak to you yet, 0:14:28.240 --> 0:14:32.040 their brain had set up in expectation and the rabbit 0:14:32.200 --> 0:14:51.400 doesn't match it. One way that we know that comedy 0:14:51.480 --> 0:14:55.920 is about breaking expectations is because comedy has to surprise. 0:14:56.640 --> 0:14:59.240 Imagine that you're at a dinner party and the guy 0:14:59.280 --> 0:15:02.000 across from you tells a good joke and everybody laughs, 0:15:02.400 --> 0:15:05.760 and then he tells exactly the same joke a second time. 0:15:06.040 --> 0:15:09.080 It would be zero funny. The second time everyone would 0:15:09.240 --> 0:15:11.880 look at one another. So what does that tell us? 0:15:12.280 --> 0:15:16.600 That something funny has to surprise. So what we've seen 0:15:16.720 --> 0:15:20.880 so far is that humor has to betray your expectations. 0:15:20.920 --> 0:15:24.160 But I've left something out so far. There's another element 0:15:24.440 --> 0:15:28.360 to something being funny, because we quite often find that 0:15:28.440 --> 0:15:31.760 our expectations don't get fulfilled, and almost all of these 0:15:31.800 --> 0:15:36.960 situations aren't actually funny. Imagine you open your suitcase and 0:15:36.960 --> 0:15:40.440 you realize that you grabbed the wrong one at the airport. 0:15:40.840 --> 0:15:44.120 That would be surprising, Or an airplane engine suddenly falls 0:15:44.160 --> 0:15:47.080 onto the roof of your car. That would certainly violate 0:15:47.120 --> 0:15:51.200 your expectations, but it wouldn't be funny at all. So 0:15:51.320 --> 0:15:56.200 comedy requires a second element. It requires a sense of harmlessness. 0:15:56.920 --> 0:16:01.440 If some event or some punchline triggers fear or anxiety 0:16:01.560 --> 0:16:05.240 or disgust or whatever instead of amusement, it's not funny. 0:16:05.720 --> 0:16:11.520 Humor requires a very delicate balance of unexpected and harmless. 0:16:11.920 --> 0:16:14.040 And by the way, I should specify that this means 0:16:14.080 --> 0:16:18.280 a sense of harmlessness for the listener. There can be 0:16:18.760 --> 0:16:22.480 cruel jokes against other groups of people, out groups, and 0:16:22.520 --> 0:16:25.400 the jokes can involve bad things happening to them, But 0:16:25.480 --> 0:16:29.360 the idea is that the listener doesn't feel threatened. So 0:16:30.080 --> 0:16:32.120 if you are the space alien and you are being 0:16:32.240 --> 0:16:35.720 quizzed about what is required for human comedy, that's the 0:16:35.760 --> 0:16:38.480 main thing. It has to surprise and it has to 0:16:38.560 --> 0:16:41.080 not threaten the listener. And we'll come back to this 0:16:41.120 --> 0:16:43.560 in a few minutes, but first I want to mention 0:16:43.680 --> 0:16:45.800 that I was thinking about this harmlessness issue the other 0:16:45.880 --> 0:16:48.360 day and I realized this is consistent with something that 0:16:48.440 --> 0:16:52.120 I read from Woody Allen many years ago. He said, 0:16:52.680 --> 0:16:59.120 comedy equals tragedy plus time. In other words, events that 0:16:59.200 --> 0:17:03.520 are pain full or upsetting can over time become sources 0:17:03.560 --> 0:17:08.760 of humor. The time gives some emotional distance, so when 0:17:08.800 --> 0:17:13.560 something tragic first occurs, the emotions around it, like the 0:17:13.720 --> 0:17:16.400 grief and the anger and the shock, they're too raw 0:17:16.560 --> 0:17:19.679 for people to find anything funny about the situation. But 0:17:19.760 --> 0:17:24.440 as time passes, emotions tend to soften, the pain goes down, 0:17:24.840 --> 0:17:27.200 and people can look back on an event with a 0:17:27.240 --> 0:17:31.840 different mindset, and sometimes they see absurdities or ironies that 0:17:31.880 --> 0:17:35.440 weren't apparent before. So you and your partner lose your 0:17:35.440 --> 0:17:38.679 tempers about something because you said X and your partner 0:17:38.680 --> 0:17:42.160 thought you meant why, and you're both furious. But six 0:17:42.200 --> 0:17:44.960 months later you both love telling the story and you 0:17:45.000 --> 0:17:47.920 get comic value from it each time. And, by the way, 0:17:47.960 --> 0:17:51.200 comedians use this all the time to craft new material. 0:17:51.240 --> 0:17:53.800 The idea is that with just the right amount of time, 0:17:54.440 --> 0:17:57.960 this sting of some tragic event can fade enough that 0:17:58.040 --> 0:18:02.040 people can laugh at it without guilty or uncomfortable. But 0:18:02.080 --> 0:18:05.359 the time component is crucial because the comedian has to 0:18:05.400 --> 0:18:09.119 be sure that the audience is ready to find humor 0:18:09.160 --> 0:18:12.160 in the tragedy. In any case, time works because sometimes 0:18:12.160 --> 0:18:16.480 we have a perspective shift, we've changed how we collectively 0:18:16.560 --> 0:18:20.119 interpret the event, and sometimes it's just a coping mechanism. 0:18:20.440 --> 0:18:22.520 But in any case, this comes back to the point 0:18:22.760 --> 0:18:26.639 that humor can't be something that is directly threatening to 0:18:26.720 --> 0:18:30.000 us right now, So let's return to incongruity theory, which 0:18:30.000 --> 0:18:34.160 is that something funny has to violate your expectations. Now 0:18:34.280 --> 0:18:37.560 that's how this is thought about by philosophers. But now 0:18:37.560 --> 0:18:39.639 that neuroscience is on the scene, I think there's an 0:18:39.680 --> 0:18:42.320 even richer way to think about this. And this is 0:18:42.359 --> 0:18:45.200 something you hear me talk about on almost every episode. 0:18:45.240 --> 0:18:49.280 This is the issue of the internal model. So in 0:18:49.320 --> 0:18:53.280 the darkness of your skull, you have a model of 0:18:53.359 --> 0:18:55.560 what the world is and how it's working, and this 0:18:55.600 --> 0:18:58.399 is what gives you an understanding of what's going on 0:18:58.520 --> 0:19:01.960 out there, and it gives you your pre So imagine 0:19:01.960 --> 0:19:05.879 you hear a comedian tell his story. The comedian says, 0:19:06.320 --> 0:19:08.959 so I walked into the doctor's office and the doctor 0:19:08.960 --> 0:19:11.159 says to me, take off your clothes and put them 0:19:11.200 --> 0:19:14.919 there in the corner. Next to mine. Now, what the 0:19:14.920 --> 0:19:18.800 comedian's doing here is taking advantage of the assumptions of 0:19:18.880 --> 0:19:23.240 your internal model, the thoughts that went unthunk, even though 0:19:23.240 --> 0:19:26.080 they certainly could have been true that the doctor was naked, 0:19:26.320 --> 0:19:29.720 but they were never even considered until you were told 0:19:29.760 --> 0:19:34.440 that information, and then you have to retrospectively revise your 0:19:34.480 --> 0:19:37.399 whole model of that situation. Or take the joke I 0:19:37.480 --> 0:19:40.600 mentioned earlier about how my dog used to chase people 0:19:40.600 --> 0:19:44.240 on bikes a lot. Linguistically, this could have been interpreted 0:19:44.280 --> 0:19:47.040 by your brain as the dog riding the bike, and 0:19:47.119 --> 0:19:50.639 possibly some neurons in your brain entertain that relationship for 0:19:50.720 --> 0:19:53.239 a fraction of a second, but then they dismiss it, 0:19:53.600 --> 0:19:57.919 and so that unconscious possibility never even gets close to 0:19:57.960 --> 0:20:01.200 your consciousness until you suddenly have to do a wholesale 0:20:01.680 --> 0:20:05.320 go back to the beginning and re understand the whole situation. 0:20:05.640 --> 0:20:08.960 And by the way, while we're talking about subverting expectations, 0:20:09.000 --> 0:20:11.640 this is not just used by comedians, but also by 0:20:12.119 --> 0:20:16.080 mystery novelists and magicians. They all have something in common. 0:20:16.600 --> 0:20:22.320 They rely on surprising your internal model to entertain. So 0:20:22.480 --> 0:20:25.800 for the mystery novelists, it's the plot twist They set 0:20:26.080 --> 0:20:28.840 everything up for you, and if they're doing their job right, 0:20:28.920 --> 0:20:32.560 they know that all these clues can be scattered around you. 0:20:32.960 --> 0:20:35.600 But right now you think a particular model of the 0:20:35.640 --> 0:20:39.440 world is true, and then they surprise you by showing aha, 0:20:39.520 --> 0:20:43.040 all these things you saw are consistent with a different 0:20:43.119 --> 0:20:47.040 structure than the one you assumed, and with the magician, 0:20:47.119 --> 0:20:49.840 it's the same thing. They move their hands, they show 0:20:49.880 --> 0:20:52.240 you this, they draw your attention to that, and all 0:20:52.280 --> 0:20:56.080 along they are making sure that you believe a particular 0:20:56.240 --> 0:20:59.399 model of the world. They need you to believe that 0:20:59.480 --> 0:21:03.639 everything is operating one way, and then they violate your expectation. 0:21:03.760 --> 0:21:08.280 They flout your brain's predictions and leave you reveling in 0:21:08.320 --> 0:21:12.080 the mystery that one thing seemed true, but you're now 0:21:12.080 --> 0:21:14.720 looking at a different situation. And if you were that 0:21:15.160 --> 0:21:18.480 space alien trying to figure out humans, you'd probably notice 0:21:18.480 --> 0:21:21.119 that we spend a lot of our time flocking to 0:21:21.640 --> 0:21:26.760 mystery novelists and magicians and comedians. You would fly back 0:21:26.800 --> 0:21:29.400 to your planet and try to explain to your alien 0:21:29.480 --> 0:21:34.600 colleagues that humans continuously pay money to have these professionals 0:21:34.840 --> 0:21:39.240 give us some sort of story and then subvert our expectations. 0:21:39.480 --> 0:21:43.720 So why do we find this appealing. One possibility is that, 0:21:44.080 --> 0:21:47.280 as we're always trying to build a model of the world, 0:21:47.840 --> 0:21:52.479 we derive pleasure from the discovery of a flaw in 0:21:52.560 --> 0:21:57.320 our model. In other words, we are machines built to 0:21:57.440 --> 0:22:00.879 self correct, and it is rewarding to find an error 0:22:00.920 --> 0:22:03.800 and fix it. So part of the joy of humor 0:22:03.920 --> 0:22:06.960 is the feeling that you had a model and now 0:22:07.000 --> 0:22:09.440 you have a better one, a model that has more 0:22:09.840 --> 0:22:14.000 explanatory power. So in the three guys walk into a 0:22:14.040 --> 0:22:17.000 bar joke, the first guy sets an example, the second 0:22:17.040 --> 0:22:21.119 guy gets your internal model to believe that it understands 0:22:21.119 --> 0:22:24.119 a pattern, and then the third guy reveals there's a 0:22:24.600 --> 0:22:28.000 different way, a broader way that you could have understood 0:22:28.240 --> 0:22:33.520 the whole premise, I suggest we find it rewarding to 0:22:33.680 --> 0:22:36.960 change the internal model, and this is what solving riddles 0:22:37.040 --> 0:22:41.120 is about, or why people listen to science podcasts. There 0:22:41.200 --> 0:22:44.840 is some pleasure derived from the realization that your model 0:22:45.000 --> 0:22:49.200 was incomplete and now you see things with a broader 0:22:49.280 --> 0:22:51.280 view than you were able to just a moment ago. 0:22:51.640 --> 0:22:54.840 At its heart, I suggest that comedy is all about 0:22:54.920 --> 0:22:58.040 thinking that you have the world understood and then finding 0:22:58.119 --> 0:23:01.560 you didn't and you now have of a bigger model, 0:23:01.800 --> 0:23:05.480 and again when the change is harmless, not threatening, then 0:23:05.680 --> 0:23:08.040 we find it funny. Now, I want to take a 0:23:08.080 --> 0:23:10.520 one minute tangent here to tackle a question that I've 0:23:10.560 --> 0:23:15.600 been wondering about for a while. Can artificial intelligence be funny? 0:23:15.800 --> 0:23:18.120 And the answer to this, I think is very illuminating. 0:23:18.760 --> 0:23:22.080 On the one hand, AI can sometimes be hilarious and 0:23:22.119 --> 0:23:24.919 tell you great jokes, but this is because it is 0:23:25.000 --> 0:23:28.800 repeating or modifying a joke that has been said somewhere 0:23:28.920 --> 0:23:31.920 in the vast corpus of human writing, even if you've 0:23:31.960 --> 0:23:34.639 never heard it. A couple of episodes ago, I named 0:23:34.680 --> 0:23:38.680 this the intelligence echo illusion, by which I simply mean 0:23:38.720 --> 0:23:42.000 that we often mistake the echo of things that other 0:23:42.040 --> 0:23:47.040 people have said before for intelligence in AI, or in 0:23:47.080 --> 0:23:50.480 this case, for true comedy in AI. But it's nothing 0:23:50.520 --> 0:23:53.440 but repeat in any case. Although AI can repeat a 0:23:53.480 --> 0:23:56.960 good joke, what AI stinks at is making up an 0:23:57.200 --> 0:24:00.960 entirely new joke. I did some experiment to drill down 0:24:00.960 --> 0:24:04.320 on this. It's not because AI can't figure out what 0:24:04.359 --> 0:24:08.400 people expect next and then shift away from that. Instead, 0:24:08.400 --> 0:24:11.360 it's because a good joke has to replace the expectation 0:24:11.840 --> 0:24:15.520 with an alternative that also makes sense with a new frame. 0:24:16.040 --> 0:24:20.240 The punchline can't just be unexpected. Instead, it has to 0:24:20.280 --> 0:24:25.880 be cleverly engineered to have some second meaning, a second interpretation. 0:24:26.520 --> 0:24:28.600 So I'll give you an example that I just ran 0:24:28.640 --> 0:24:31.840 this morning. I didn't want chat gpt to tell me 0:24:31.920 --> 0:24:33.960 a joke i'd heard before, so I forced it to 0:24:34.000 --> 0:24:36.439 make up something new by prompting it to tell me 0:24:36.480 --> 0:24:40.280 a joke about lobsters and peanut butter. So here's what 0:24:40.280 --> 0:24:41.000 it came up with. 0:24:41.280 --> 0:24:44.720 Three lobsters walk into a grocery store. The first lobster 0:24:44.800 --> 0:24:47.520 goes to the seafood section and asks for butter. The 0:24:47.560 --> 0:24:50.000 grocer gives him a tub of butter, and he happily 0:24:50.040 --> 0:24:53.359 scuttles away. The second lobster goes to the same section 0:24:53.480 --> 0:24:56.200 and asks for butter. The grocer gives him a tub 0:24:56.240 --> 0:24:59.920 of butter, and he too happily scuttles away. The third 0:25:00.480 --> 0:25:03.399 goes to the grocery stores pantry section and asks for 0:25:03.480 --> 0:25:08.200 peanut butter. The grocer looks surprised and asks why peanut butter. 0:25:08.600 --> 0:25:10.960 Aren't you going to cook with butter like the other lobsters. 0:25:11.359 --> 0:25:15.199 The third lobster replies, Nah, I'm allergic to butter, I 0:25:15.240 --> 0:25:17.160 always shell out for peanut butter instead. 0:25:17.760 --> 0:25:20.920 What now, I would call that the dumbest joke I've 0:25:20.920 --> 0:25:24.119 ever heard, But it doesn't even qualify as a joke. 0:25:24.400 --> 0:25:28.959 Why because the punchline, while it is a violation of expectation, 0:25:29.640 --> 0:25:34.160 doesn't offer a second interpretation that makes sense. So crafting 0:25:34.200 --> 0:25:36.680