2. Evolution of Behavior I | New Lecture Series 2024 | Robert Sapolsky | Human Behavioral Biology | Stanford

Setting the Stage for Human Behavioral Biology OK, let's begin. Just as a heads up, almost certainly not today, but on Friday we'll be talking some aspects of the evolution of infanticide if that's going to be something that's troubling. OK. So starting off, what we now know is basics of Darwin and basics of evolution, and we know what things are not the case about evolution and we know evolution is for real and all of that. We know something about natural selection versus sexual selection, balanced selection, frequency dependent. What this has poised us for right now is to see how all of this applies to behavior. And the whole underpinning of it is exactly as with the skulls the other day, there's an underlying logic, just as there's a logic to Physiology of kidneys and desert rats and all of that. There's a logic to behavior as well in an evolutionary context. So what we start off with is what's going to be in lots of ways, the central question for a number of lectures to come, which is, OK, you got a behavior and you're speculating about how it evolved. And by definition, if you were talking about evolved, you have just committed yourself to. And genes have something to do with it. They may not be causative, and we will see. Genes are rarely causative, but they play a role. And thus, if we're talking evolution, we're talking genes. At which point you got to ask the question, where are the genes? Show me the genes, How are they working? All of that, and much of what the next 345 lectures are going to be about is how do you tell when there's a genetic influence on a behavior? From an evolutionary perspective, from a molecular biology perspective, that is where we are heading. OK, so we start off from the standpoint of evolution of behavior. There are three main pillars that explain everything and don't write them down yet because we'll get to them in detail, but individual selection, kin selection, and reciprocal altruism. OK, starting off with individual selection, beginning a caveat throughout this, we're going to be talking about what a vampire bat wants to be doing with her babies. We'll be talking about what one tree would like to do with respect to nearby trees, that it's related to all of that. This is just figures of speech. This is just a way of saying over the course of evolution, vampire bat mothers who did X left more copies of their genes. It's just a shorthand where we're going to be using. What would this Brian shrimp want to be doing in the circumstance? So we start off with a scenario, and this was a scenario that I sought tons of when I was a kid watching all these wildlife programs. There was this one called Mutual Mutual of Omaha's, Oh, what's it called? Wild Kingdom, something like that. And it was totally great and amazing and watching it. And for years, it was sponsored by Mutual of Omaha Insurance Company so that you always had these weird, really awkward transitions before a commercial where the guy would have to say, you know, just as rhinos will mate for hours on end, you want home insurance for your valuable home and that kind of thing. Now, the host of Wild Kingdom was this guy named Marlon Perkins. And we start off with him because Marlon Perkins taught all of America about evolution. And the trouble is, he taught America totally incorrectly. Here's what the scenario would be. You're sitting there watching, and it's about the wildebeest migration in the Serengeti. Throughout the year, wildebeest follow this cyclical pattern of rains. And this herd of one and a half million wildebeest move around, and it's like one of the most magnificent things imaginable. And so you're the wildebeest, and there's always more fresh grass ahead of you as long as you all keep on the move. And this whole bunch of wildebeests, however, have reached a problem. They have come up to the end of this field that they have just eaten down to their nubbins now of all the grass. And ahead of them is a great field it, but there's a river in between. There's a river in between which very inconveniently is filled with crocodiles who were waiting to nab wildebeest and feast on them. And then the wildebeest have this challenge now of how do you get across the river to get to the other side and avoid the crocodiles. And they're all standing there like 100,000 wildebeest on the edge. And they're Hemming and hawing and pooping and all agitated as to what to do. How are they going to get across the river? And suddenly, suddenly, from the back of the crowd of wildebeest comes the solution. This elderly wildebeest forces his way through the crowd, comes up to the edge of the river, says I sacrificed myself for you, my kinder, and throws himself on the river. And all of the crocodiles fall on him. And while they're busy with him, the 100,000 wildebeest come down, swim across and come out the other side. Oh my God, that is amazing. That's amazing. And Marlon Perkins, as we watched this amazing thing, would then say, ask the question we were all wondering, why did that elderly wildebeest sacrifice himself? And where Marlon Perkins taught all of America incorrectly was an inevitable sound bite, which was the animals behave for the good of the species. The elderly male wildebeest knew that this was by doing the sacrificial thing, everyone was going to flourish. Animal behavior is for the good of the species. And that turned out to be gibberish. And it was only with more observation, say of wildebeest in that traumatic setting, only with more observation that you would see what's actually going on, which was this elderly wildebeest was not fighting his way to the front to sacrifice himself. He was being pushed forward by everybody else because he was weaker than them. He couldn't fight. They pushed him and then they pushed him into the river. I sacrifice myself for you, the species my ass. They pushed him in because he was the weakest and he got dumped in there. And at that point the Crocs got busy and everybody moved on. And what we have this transition from was this dominating view of Marlon Perkins and all the serious evolutionary biologists thinking about behavior, what they called group selection, group selection, Why do animals behave? They behave for the good of the group, the good of the population, the good of the species. And this completely dominated thinking at the time. And everybody got raised on this. And it was not until like late 60s, early 70s that people began to realize that this group selection stuff made no sense at all because you would keep seeing animals behaving in ways that were not for the good of the group. It was them behaving for the good of themselves. The Selfish Gene and Individual Reproductive Success And this slowly ushered in the very first sort of foundational pillar of the whole field, which was people saying group selection is nonsense. Selection is not on the level of groups. Selection is on the level of does this individual pass on copies of their genes to the next generation. And as such, this brought forward this first megalith of explaining everything, what was termed individual selection. Individual selection. Richard Dawkins, Oxford zoologist, writer extraordinaire, coined the term the selfish gene to catch this notion. Animals are not behaving for the group, they are behaving for leaving as many copies of their genes as possible in the next generation. What we'll see in a while is the selfish gene is not quite accurate. It should be the selfish Chino. What we'll also see in a while is after, thank God, group selection was trashed in the 60s, it comes back later in a form that's for real and so much more interesting. OK, but meanwhile we've got individual selection. Animals are behaving for their own selfish needs to maximize the number of copies of genes they are leaving in the next generation. And what you could begin to view like animal behavior about is all of that is just to get your genes into the next generation. Remember, reproduction of the fittest, not survival of the fittest. All of this is about this. And this gave rise to this great quote, this guy Samuel Butler, who is kind of Darwin's time. He said sometimes a chicken is just an egg's way of making another egg. Sometimes all this behavior stuff and there's complications and competition and unrequited whatever, sometimes all this behavior stuff is just to get another egg, another end to the next generation. Sometimes you could think of all of animal behavior as just animals trying to maximize their own reproductive success, maximize the number of copies of their genes they are leaving in the next generation. So with this mindset, you look at animal behavior and yeah, the old wildebeest wasn't jumping in there voluntarily. And what you see is this whole world of aggression and competition and all of that, and animals striving and salmon leaping over dams, all to leave copies of their own genes in the next generation and to maximize that as much as possible. Okay, So what would individual selection look like in the context of natural selection? Obviously you run away from the predator as fast as possible. What would it look like in the context of sexual selection? Obviously as well, you do whatever traits or in vogue this season so that somebody's willing to mate with you so you could pass on copies. So our first building block. Animals do not behave for the good of the group or species. Animals behave to maximize their own individual reproductive success, individual selection, the selfish gene. Sometimes a chicken is just an egg's way of making another egg. Good. How Shared Genes Drive Altruistic Behavior So this transitions us now to the second great pillar, the second great pillar of all this stuff which is built around like this thing, this guy Mendel. And what did Gregor Mendel teach all of us if anyone paid attention to him at the time? Mendel taught us 2 critical things. Why it is that we resemble our relatives, and 2nd, why it is that nonetheless we're not exactly like our relatives. Unless you're an identical twin, and that's not even the case. And we're getting to that plenty relatedness. He figured out the mechanism of inheritance over the course of generations. He found the evidence for the existence of genes, even though the word didn't exist yet. Mental showing us. Here's how you're related to your relatives. You share genes with them. You share genes. And this critical point with Mendel, with Mendelian, all of this is the more closely related you are to someone, the more genes you share in common. How's that work? You're getting on the average half of your DNA from each parent. So you are one half related to each parent. You share half of your genes with them. As a result, if you have a full sibling, you share half your genes with a full sibling, 1/4 of your genes with a half sibling. And from there, far out from there, as I plan to attend the catch up session on mental, from there you just see how much of your genes you share with a first cousin, a second, third. And in each case there's a doubling or there's a halving of the percentage of genes you share with them. You are more closely related to closer relatives, as in you share a greater percentage of genes. OK, important digression here. Some of you will be sitting there saying, wait a second, wait a second. So you share half your genes with each of your parents. And then we sit and we read stuff like we share 98% of our DNA with chimps. What's the deal? Are we more related to chimps than we are to our parents? No, this is an apples and orange situation. It's referring to two different things. When people are saying we share 98% of our genes with chimps, we're talking about genes specified for particular traits. We and chimps have the gene for growing a nose. It's not just one gene. And it's genes for growing eyebrows. We have those genes. Chimps have them. Redwood trees do not. Crabgrass does not. On the other hand, neither we nor chimps have the genes for growing antlers. And so those are the attributes for having a symmetrical body pattern, for having two ears, having forelimbs. We share all that stuff with chimps and lots of brain stuff. We share the trait. We both possess genes relevant to that trait. When we're talking about you and your parents or you and a sibling, we already take it for granted neither of you have genes for antlers and that all of you have genes specifying you make eyeballs and they have colors and whatever sort of thing. You're already taking that for granted. What you're focusing on there is not that you share a trait, you share a particular flavor of the trait, a particular you have the gene for for the liking ice cream and all the members of your family do as well. And it happens the ice cream liking gene comes in different flavors and one parent likes vanilla, the other likes strawberry, and you wind up sharing the proclivity 50% with each parent. It's for variation in these traits for different flavors of different versions. So that's how it is that you can share 98% of your DNA coding for just broad categories of traits with chimps, 96% with monkeys, all of that while you only share half your genes, half the flavors of particular genes with each one of your parents. Good. So we got that out of the way. So thanks to individual selection, you want more than anything on Earth to leave a copy of your genes in the next generation. But now comes in this recognition that sort of underlies our second pillar here, which is if you take a full sibling who shares half your genes with you, and if instead of you reproducing once, you make it possible for them to reproduce twice, you've just left as many copies of your genes in the future generations from you. It's 1X from them. It's 2 * 1/2 X. From the standpoint of Darwin and fitness and math and all of that, it's exactly equivalent. And what that ushers in is the second category of ours, kin selection. Yes, sometimes the best way to leave lots of copies of your genes is to compete and be selfish in an individual. But sometimes it is more efficacious and more sort of fit over the course of time to help relatives reproduce. And this gave rise to one of the all time great quips. This this evolutionary biologist, Hall Dane, early in the 20th century came up with this quote. One time he was sitting at the bar, which apparently he did 24/7. He was sitting at the bar and somebody was confused about this whole issue and he like grabbed A napkin and scribbled out his calculations and made the pronouncement that explains all of kins selection. He said I would gladly lay down my life for two brothers or 8 cousins because if instead of you reproducing, you die helping them and the two brothers reproduce, it's exactly the same. If you die so that the 8 cousins can reproduce, it's exactly the same. I will gladly lay down my life for two brothers or 8 cousins. In the 1930s when he said this, sisters didn't exist yet so it was just brothers that he was thinking about. Oh, sometimes you maximize your reproductive fitness by way of individual selection. Sometimes you do so by helping relatives, by aiding relatives as a function of how related they are to you. And this was finally formalized in the 60s by this guy, WD Hamilton, who was a God in the field. And he came up with Hamilton's law, which quantifies when you should be doing what to help out a relative and how much cost should should you be willing to take on to help some relative pass on copies of your genes. And as long as it's less than the benefit. Something is spelled wrong there. As long as it's less than the benefit to this relative times their degree of relatedness such that a full sibling is going to account for a whole lot more than a 14th cousin. All of that, as long as the benefit divided by that comes out to more than the cost to you, it makes sense to help them. Hamilton's law showing quantitatively. Here's when you should help a relative when you get more out of it by helping them as a function of how related they are to you than just focusing on your own striving towards enhanced reproductive success. Now what this did was usher in this whole universe of. Of making sense of animal behavior, including us, which is that species after species, social species of all sorts out there are obsessed with kinship, just like humans. Where, like, as far as I can tell, half of what anthropologists do is find out in the group they've been hanging out with. Like what do you call the sister of the aunt of your uncle? Kinship terms Animals the world over are completely obsessed with kinship. So where do you see kin selection? Where do you see organisms behaving to maximize their own reproductive success by way of helping relatives? And, and you see it everywhere. You see it with bacterial strains where you will get more cooperation among bacterial lines that are closely related than ones that are not trees. Tree behavior. I don't even know if trees behave. I know there's people who insist they do, but you've got a tree. You are a tree, You're a tree and you've got roots out there and nitrogen or nutrients or something that you could or could not share with the next tree to be helpful, to be cooperative with them. And what you see is trees share their nutrient good stuff at the end of their root stuff. Trees share it as a function of the relatedness of the next tree along. If it's a total stranger, don't do it. If it's a close relative, do it. What's this about? It's tree species where they dropped their seeds. So the tree next to you is likely to be a relative versus ones where like some bird flies away with your seed miles and miles and drops it down there. Are you in a forest of trees that are closely related and you see this new tree and sharing going on there? How does a tree know that it's related to another tree? How does a bacteria know that? How do we know that? We're going to have a whole lecture early next week on how you recognize relatives. Vervet monkeys live in East Africa. They're kind of interesting. They're way inferior to bad wounds. However, this was work done by by Dorothy Cheney and Robert Seyfarth, University of Pennsylvania, who pioneered this whole thing of doing experiments out in the field with your monkeys. What you do first is you make recordings of every vocalization, type of vocalization given by every member of the group you were studying. So you've got this library of alarm calls from every individual, solicitous calls. I'm happy vocalizations. Well, then you've got this whole library. And what they would do then is they'd put a speaker in the middle of some bushes and they'd play a recording of somebody giving a particular vocalization and then watching to see what everybody does. Now, what they would show with this, Ah, First off, you're saying what this requires is that vervet monkeys can recognize individual voices. Yes, indeed, absolutely. All primates can do that. So you've stuck in the bushes there, your speaker, and you play a recording of infant A giving an alarm call. Giving an alarm call. And what does everybody else do? They all look at Infant A's mother to see what she's going to do. Now they understand relatedness. They understand if an infant is giving an alarm call, who's the individual who has the biggest incentive in paying attention to this? And everybody immediately looks at, Oh my God, that's like Madge's kid. Does she hear that? And what are you going to do now? What next? They understand relatedness along those lines. Next example of showing vervet monkeys, just as an exemplar here of understanding subtle stuff about kinship. OK, so you've got female monkey A&B and their kids, Little A and Little B. And it happens. Little A is a jerk and beats up on Little B, observed by B's mother and at a higher than chance rate. Later in the day, Mother B is going to pummel Mother A. Whoa. She understands. Not only is this my kid, she understands my kid was just treated badly by the child of this female who I could dominate and their understanding relatedness there. Along those lines, all you need to do is like sit in the middle of a baboon troop and there's like 80 animals scattered around and a lion suddenly pops up and what do you see? Within 3 seconds, every single female who has a kid sprints across the field to grab her kid and go flying up a tree. At that point, kin selection in action. Natural selection. You were helping out this individual who shares half their genes with you. Makes perfect sense. Green Beards: Cooperation Through Shared Traits Next layer, which is reciprocal altruism. Sometimes it makes sense to cooperate with another individual even if they are not irrelative. And this is going to put us into very complicated terrain. Before that, it turns out there's kind of an intermediate form between kin selection and reciprocal altruism. Kin selection, I help them if they share a lot of genes with me. Reciprocal altruism, under the right conditions, I will cooperate with someone even if they don't share any genes with me. An in between form called green beards. Green beards The term was supplied by Richard Dawkins yet again with his selfish gene green beards. What is a green beard trait? And he and others first speculated that these things exist. A green beard trait is when you're only having one trait per ( 1 gene. If you go with our simplistic assumptions that we're going to trash over the next few weeks, if you share 1 gene with another individual under a special circumstance, what would be a green beard gene? What it does is it generates A conspicuous signal from you, your green beard. What it does is generate the same exact thing in anybody else in the population who happens to have that gene. So first requirement is you display your green beardness. The second sort of component of it is you could recognize other green beards. You're in there and there's 100,000 wildebeest and there's one back there with the green beard and you're a green beard. You can spot each other. You could find each other and 3rd attribute, you are cooperative with somebody who has a green beard, and what you're getting here is some sort of poor man's kin selection. It's not sharing half of your genes, all of that. It's just if you see someone who shares a conspicuous trait with you, cooperate with them and this winds up being something that actually occurs. People went for years saying in principle you could get green beard effects like this, but no one can ever find any. Here's an example of a green beard effect. You are a sperm and you are very intent to getting to your target. That egg there and you happen not to be a terribly good swimmer and things are not going well here. We have two green beard sperm. Each of them has a distinctive protein on its surface. No other sperm have one in 100,000. But you've got this distinctive protein and you're able to recognize somebody else with that distinctive protein. How do you recognize that the two versions of the protein dock together? You know you have found another green beard sperm. If the two of you can dock together and what happens then? Two sperm put their tails together and they swim faster. We will see all sorts of ways in which you get aggregates of sperm that collaborate and then sperm that try to poison each other that compete. But in this case, these are two green beard sperm who because of this very rare trait that recognizes each other and now they are able to altruistically, effectively swim faster. So that's one of the first examples of green beard traits popping up in the real world. If you allow for a certain metaphorical thinking, some of the most interesting things out there are green beard traits where we're not really talking about genes and evolution when we're talking about cultural gene beard traits. And I happen to know for a fact that Richard Dawkins hates when people start talking about green beards metaphorically as cultural memes, But tough luck, because they're incredibly interesting. What am I talking about? Like a cultural green beard. You're standing there and there's a big crowd of many, many people and you're wearing a Stetson hat. A Stetson hat because you've just rode in from the planes and the deers in the Antelope. You are wearing a Stetson. And if you look through the crowd and you spot somebody else with a Stetson, you have a much higher than chance level of sharing a trait with them. Both of you think that cows are good things to eat, and maybe you'll even cooperate to hunt down a cow to eat. On the other hand, if you're wearing a sari and you spot somebody in the crowd there, you know that you share the trait with them that cows are not for eating, cows are for worshipping. Just by seeing that singular, the hat you're wearing, the type of clothing, you already know this is an us and not a them, and you're getting all sorts of cultural in greed out greed stuff resting on a single attribute like that. Other versions of it, you go around and you see certain types of religiously devout men, and depending on the particular covering they have on their head, you already know which one thinks which religion should have control of the Temple Mount in Jerusalem just by all the signal stuff. How can you look across a crowd and say, whoa, they're one of us, I'd recognize them from anywhere, let me go find them and the two of us are going to work together. Cultural red beards are incredibly important because by the time we get the stuff about aggression and cooperation, what we see is we can have multiple green beard categories in our heads and which one is counting as most important can change in an instant, can change as to who counts as an us and to them. The Logic of Cooperation and Cheating in Reciprocal Altruism But what we now move to is our third pillar, our third pillar, which is this business about reciprocal altruism. Sometimes it makes sense to cooperate with an individual who's totally unrelated team shares no genes in common. It makes sense to cooperate with them if and only if it's to your advantage they reciprocate. And by having this reciprocal, helpful altruistic relationship, things work better than if either of you were doing it on your own. This is the whole logic behind reciprocal altruism. Now note that this is all within this backdrop of competition. And you want to leave copies of your genes or helping close relatives to leave more copies than them and them. And this is suddenly in the world in which, like, cooperation is possible instead. And we're going to spend a ton of time seeing, how do you jump start cooperative reciprocal altruistic relationships? Now, people in the field thought they had found a great version of reciprocal altruism, but when they looked closely, they saw it wasn't really the case. Guy named Brendan Bohannon, who was in the department here some years ago before he moved to the Pacific Northwest. And what he studied were three different strains of E coli bacteria, and they had different attributes. This strain is able to make a poison. It's expensive for it, though it's metabolically costly. It makes this poison, which it could release. The second strain is sensitive to the poison. It makes it less healthy. However, it's very good at taking up nutrients and getting lots of food. The third strain is not poisoned by the poison, but it doesn't take up as much food. So strength, weakness, strength, weakness, strength, weakness. And what he showed is you put three different strains like this in the right proximity to each other and they form a rock, paper, scissors relationship. Rock, paper, scissors. It's exactly like Charmender whose fire and Bulbasaur whose grass, and Squirtle whose water. Thank you to my kids for writing this out for me. Each time you get a rock, paper scissors. This one can drive them into extinction. This one can drive them into extinction by being much better out competing them for food. This one can drive them into extinction because it doesn't have to spend money on making the poison. And what you see is the logic of a rock paper scissors scenario. Oh, if you are the poisonous one, would you want to drive these all the way to extinction? No, because if they all die off, you're going to have a gazillion of these that are going to outcompete you. Each one of these evolves an incentive for constraint there only partially trying to do in the other because every time you completely wipe out your victim, the predator who trounces you is going to swarm into large numbers. So people finding some examples of rock paper scissors like that with bacterial strains things of that sort is this piece is this the garden of Eden of oh constraint from maximal aggression instead of that you only a little bit mean and mean all of that is this no, this is not cooperation. This is like rock, paper, scissors is to like actual cooperation as like mutually assured destruction strategies with nuclear missiles is to World Peace. It's not about peace. It's just you really don't want to do that because it's going to come back to bite you in the rear. You want to have constraint. So you have these systems there that people initially mistook for utopia. And no, it's just to balance scenarios of stuff. Actual, actual cooperation is something much rarer. And what we see with reciprocal altruism is that critical requirement. You help out this other individual and you do so with a high likelihood of them reciprocating back. So you see versions of this all over the place. What was the initial notion who, what species would show reciprocal altruism? And it was obvious which would have to be, you got to be smart. You got to be smart so you could recognize individuals. So, you know, this is the one who owes me a favor back. You need to have social species that are in stable groups and are smart and recognize individuals, all of that. That's where you see reciprocal altruism. But then people began to see, you actually see it all over the place. There's this whole new field that people are calling socio virology, viruses that cooperate with each other in reciprocally altruistic ways. Two different viral strains where one of them is very good at getting into a cell and and pulls the virus and the other virus in with it. And the other one is really good at bursting the cell when it's made lots of copies and spreading out there. And they cooperate, they work together on that. And they're neither smart nor socially stable nor having kinship terms or whatever. Nonetheless, most of what you see in the realm of reciprocal altruism are complex social species. And the logic of that is obvious in all of those, which is, you know, if you've got 2 individuals cooperatively hunting together, their chances of getting a prey item is better than either one alone. And you see that rhesus monkeys that cooperate and looking for food, each individual winds up getting more food than if they did it on their own. Hunter gatherers. Cooperative hunting greatly increases likelihood of. Yeah, many hands. What's the cliche? Many hands make the task light. Whatever. Yeah, cooperating is really good. You're a virus and you're really good at getting out of cells, but you're crappy at getting into it. And Greg, you have found another virus with the exact opposite skills, and the two of you are now Blood Brothers to the death. And you're working together, and each one of you gets more benefits out of it than if you were working alone in the 1st place. What do you want to do? You want to find someone who's cooperating with you, and what do you really want to do? You want to find somebody who's cooperating with you, who you can cheat against, who you can take advantage from them and not reciprocate. And thus, what's the other thing you really, really want to do? You want to be good at spotting A cheater and being able to punish them. And what this brings in is the whole possibility of yes, yes, cooperation is great, but if I can cheat and they continue to cooperate, I get a bigger payoff. You suddenly have the possibilities of cheating and surveillance against the possibilities of cheating. And you look at all these reciprocal altruistic systems and what you see is every now and then one of the participants cheats here. Let's look at an example of this again, that this stuff is playing out with single cell organisms, bacteria. So there's some kind of bacteria and they all live single cell lives and every now and then they. All come together because it's seasoned to reproduce and they form these things called fruiting bodies. And the bacteria make up this whole thing, the top part and the stalk. And it's only the top part that releases new cells, new bacteria, thanks to cell division. So you've got this altruism problem here, which is you don't get the stability for this to form unless you've got this stalk sort of holding everything together. So in a world of perfect reciprocal altruism, you've got two different strains of bacteria. And what do they do? They cooperate. They play square and fair with each other, and half of the bacteria in this come from one of the strains, half from the other, and half in the stock come from that. They're equally sharing the advantages and the cost. And what would sheeting look like if you have some bacterial strain that gets in there early and tries to 1st monopolize this and sticks? This poor schnook strain there is stuck with just being the stalk, and this one gets to reproduce and this one does not. This is what cheating would look like. And what you see is different strains of bacteria have propensities towards trying to center themselves up here, trying to violate the 5050 social contract, and instead try to disproportionately grab this stuff. And what's the defense against that? Other bacterial strains recognize, recognize we're back to the world of recognizing another sperm sharing a trait with you. They begin to recognize it, and they withdraw docking proteins on their surface that would otherwise dock with this other bacteria. They just dock to themselves. They exclude the cheater who is not able to dock to them. So you've got bacteria cheating and you've got bacterial defenses against cheaters. This is totally amazing. And what we just saw is bacteria put a lot of effort into being on guard against cheaters. And as it turns out, so do we. So do we. There's this great psychology test where you sit them down, you give them this whole convoluted story. There's a king and the subject appears before the king. And the king says, if you do this very complicated thing and succeed, I will give you this reward. But if you fail to do that, I will cut your head off. And this whole complicated story then ensues. And at the end you find out like the person reappears in front of the king and says this is what happened. And at which point the king on the rewards that were cuts their head off. And you, plowing your way through this complex story, have to figure out was just as served the individual that they didn't actually turn out to be the way the king promised. And it turns out there's two ways in which justice could not be served. One is you do the task. You do exactly what you were supposed to do, and you come back and you present it. And the king does not reciprocate. The king cuts your head off. The king is mean to you when they're supposed to be rewarding. The alternative is you fail dismally and you come back and you confess to that and the king rewards you. The king rewards you when you have not earned it. By the laws of this story. You have these two versions where you were violating the rules, false positives and false negatives, where someone has done their task by but you cheat them and don't give them the reward. Or where somebody hasn't done their task and you reward them from out of nowhere. And what you see is with these convoluted versions of these stories, people are much, much better at spotting when the king should have given the person the reward and instead cuts their head off. People are much better at working through the logic of that than when you have a king who gives a reward to somebody who doesn't deserve it. Our cognitive structure is such that we are way, way better at spotting norm violations, norm exceptions that go in the direction of cheating rather than unexpected generosity. Chimps. Chimps have the same proclivity. They are better at spotting somebody cheating then it's spotting somebody who's being pathologically generous and being cooperative for no reason. There's been selection for this, and this is the sort of thinking that came in in the 70s or 80s. Mostly the work of this guy, Robert Trivers, big theoretician in the field, sort of worked out the mathematical logic of reciprocal altruism. And this was this revolution. Because up until then, ever since Darwin, up until that time, what is evolution about? What is evolutionary success about? It's about competing and winning and dominating an aggression. And suddenly people realize, oh, sometimes it pays off to cooperate. Sometimes evolution has selected for peace and cooperation rather than savage tooth and claw 24/7 kind of thing. This was revolutionary when this came through and this up into the whole field and completely changed people's thinking about the universality and the inevitability of aggression. And this was totally great because this was this completely new outlook, except it turned out this Russian scientist had gotten there 80 years before everybody else. The scientist named Peter Kripatkin, who I think is one of the coolest humans who has ever lived. He was born a Prince, a czarist Prince, and he renounced his royal state. He was a zoologist and geographer and writer and all of that. And he was also a revolutionary anarchist who was jailed by the czar, his like fourth cousin, any number of times for him fermenting revolution. When the revolution finally came in 1918, he was one of Lenin's buddies. But because he was an anarchist rather than Lennon's camp, Lennon conspired to have murdered. But fortunately, Kropotkin was on his deathbed by then, so Lennon didn't waste the effort on that. Kropotkin. Kropotkin, 80 years earlier than all of this, published a book whose name I am losing about mutual aid, mutual aid, the driving force of evolution. And before any of this stuff, 80 years earlier, he was already saying, you know, sometimes cooperation is good and sometimes you will see evolution selecting for that, and it tends to be in decentralized systems, says the anarchist cryptoc and all of that. None the less, the rest of the world catches up 80 years later, and in comes this recognition that under very special circumstances it makes sense to cooperate. Game Theory and the Prisoner's Dilemma So now we get the critical question that comes up, which is so when do you cooperate and when do you cheat? How do you know what the optimal strategy is? And this is when there was a very unlikely marriage that went on between zoologists and game theorists, game theorists, war strategists, diplomacy strategists. People in like diplomacy school have been being taught about game theory all along. It's very formalized math of what is the optimal strategy in different types of competitive games that have possibilities of corporate, when do you cooperate? When do you cheat? What optimizes it? And they have been like studying that for years in terms of like, OK, in what circumstances should you let loose all of the atomic weapons? And when is that going to pay off for you? Wonderfully, there's all sorts of people who made their careers as weapons strategists in the Pentagon using game theory. Very interesting guy. Daniel Ellsberg, who in the 1960s was famed and notorious. He was a person. If you know your history, he stole a whole bunch of papers from the Pentagon where he worked, a bunch of papers basically showing that you US having fabricated all of its rationales for entering into the Vietnam War. He gave it to the New York Times and gigantic thing that eventually the Supreme Court had to say it was OK for them to publish it. And this partially brought down Nixon's presidency and all of that. Daniel Ellsberg, before he did this, was a hawk war strategist in the Pentagon coming up with all sorts of things he has credited with coming up with a particular game theory strategy called the Masada strategy. Any of you who know your Middle East miseries at this point? Masada was a citadel up on the top of a plateau at the time that Israel was conquered by Rome and 60AD. And there was a whole bunch of rebels who wound up on their as their last stronghold. And as the Romans were coming, they made the decision to do mass suicide rather than being taken. And what Ellsberg came up with was this Masada strategy, which is a little bit different from what actually happened, which is it is known to everybody that if any state hostile to Israel ever overruns Israel, Israel is going to set off its nuclear weapons and take out everyone. Mutually assured destruction in this case. And Ellsberg's paper on this was called something like on the benefits of madness. That is a strategy. Make the other side think you are so damn crazy that if you had, if you're going to take out everyone, they don't care or they just want to. And this was a very influential paper, apparently, in the Israeli military. OK, so Ellsberg leaks the Pentagon Papers, and suddenly everybody discovers game theory. Oh, all these military strategists have been studying game theory as to when you cheat and when you cooperate. And the poster child for game theory and games, sort of the fruit fly of all of that, is a game called The Prisoner's Dilemma. And this was the centerpiece of all of the textbook work on when do you cheat and when do you cooperate. OK, so you've got 2 individuals, 2 individuals, and there's a whole story built around it where they do a crime together and the police get them and they're questioning them individually. And do you both promise not to rat on the other person or do you rat on them? And what's the other person going to do? You've got 4 possible outcomes. You as one of the two participants can be mean, can stab the other guy in the back or be cooperative. Stick with your agreement with them. Just as you arrested you both say, OK, I'm not going to say anything. Do you stick with that or do you stab them in the back? And meanwhile the other individual has the exact same options. So we look at what the outcome is. Both of you. Both of you are nice, both of you cooperate, you refuse to give evidence on the other. And what can the police do? You only get a medium jail sentence of 2 units of jail sentence. Okay, so that's nice. In contrast, if you both stab each other in the back, you only get one unit of reward, a much longer jail sentence. So nice and nice together. You get 2 units of reward. You're both mean, you get punished for that. And now we see the critical two other squares in the matrix. What if this is you, This is you and they're being nice and you stab him in the back. They're saying, Oh no, I didn't do this crime with this guy. And you're saying with the police, yeah, he was my partner and I'll give you proof if you let me off free. And what's the outcome when this guy cheats and they're playing by their cooperative rules? This guy gets 0 points, he's thrown in jail for a long time and this guy gets the best outcome. He gets to walk free. So what what this individual really wants is a circumstance where they know this guy is going to be cooperative and they can stab him in the back. So we have this matrix and what prisoner's dilemma is all about is when do you play nice and when do you stab the other individual in the back? What's the optimal play strategy in Prisoner's Dilemma? So you have one circumstance where it's absolutely clear what you should do. Suppose this is only going to happen one time. One time so you know this is never going to happen again. You only have one time that you make a choice and Oh my God, what if you make the choice to be nice while they're being mean? You are screwed. You get the worst possible outcome. You're a total sucker and in the field and it's called you get the suckers pay off. So just to make sure that doesn't happen because that's the worst possible outcome, I'm going to cheat and and the other individual cheats as well. I mean, it's sort of like a medium reward in circumstances where there's only one single round and you know it, one is selected for non cooperation every single time a single round and you're never going to see this person again. You're never going to be in this situation again. A single round pulls for cheating. Nobody cooperates. How about a circumstance where you know you were going to interact or you're going to play the prisoner's dilemma twice with this other individual? What's your strategy there? Well, you know, by the time you get to the last one, you're sure not going to be an idiot and be nice. So you're going to cheat there, and they're going to cheat, and that's already a given. So there's really only one round where you still have an option to make sense of. And what do you do there? It's the same logic as here. You cheat with two rounds, you never get cooperation. With three rounds, you never get cooperation. And you don't get cooperation in any version where you know how many rounds there are, because as long as you know when it's the final round, it makes sense to cheat in the final round. What you wind up getting is a world in which you get payouts if and only if there's an unknown future. You're going to see them again, but you don't know how many times. And as soon as you get uncertainty about how many rounds jargon in the field, as soon as there is a shadow of the future, there's sudden pressure to start cooperating. Because if you're a jerk here, you're going to be seeing them again. And maybe it's going to be enough times that they're going to wind up ahead of you because you don't know how many rounds there are. It's only when there is a future of uncertain length, the shadow of the future, that you begin to select for cooperation. So the question becomes, OK, so when you've got a future here, how much do you actually cooperate? How much should you be doing that? When should you cooperate? When should you cheat? And out of this came this grand, amazing paper. In the early 70s, WD Hamilton, who came up with Hamilton's Law, teamed up with a political scientist, this guy named Robert Axelrod, at University of Michigan. And then they both sat there and tried to figure out, well, what's the optimal play strategy in a prisoner's dilemma game with an unknown number of future things. And they had this inspired idea, which was they round up a whole bunch of their buddies, like 200 of them. Buddies who were economists and strategists and prize boxers and murderers and Nobel Peace Prize winners and a whole collection. And in each case they describe to this person, here's how you play prisoner's dilemma. What would your strategy be? And they collected everybody's strategies. And they took some ancient Neanderthal computer the size of like Delaware, and they ran a round Robin tournament running every single strategy against every other strategy to see what's the optimal strategy that comes out of it. And what turned out was this absolutely landmark observation as to what's the optimal strategy in a prisoner's dilemma game. They got back these incredibly complex algorithms from different mathematicians. And if and only if you do this with this probability, and what was amazing, what was legendary, what floored everyone, is the strategy that wound up with the most points at the end. The most adaptive strategy, the most successful one was the simplest 1 you could come up with, and it wound up being called tit for tat. You start off cooperating in the first round and your A, you're B, you're B, and you both cooperate, and that's great. And if you both cooperate, you do the same thing again in the next round and you both cooperate and that's great and you're getting the payoff in the matrix there. But suddenly out of nowhere, that son of a bitch A, in the next round, they cheat on you. They cheat and they've just gained an advantage. And what the strategy says is, if they've cheated against you once and then they go back to cooperating, punish them back once and then go back to cooperating. If they keep cheating, you keep punishing them back. In other words, with tit for tat, you simply do whatever it is the other player did in the previous round and this turned out to out compete every other strategy out there because it was optimal. It was optimal in four different ways. First off, your starting point is cooperation. So that's already like good cheery chances of things working out happily at the end. You start off with cooperation. You were able to punish somebody if they cheat against you. You were able to forgive them and reestablish cooperation. And most of all, it's not one of these crazy algorithms and probabilities. It's clear. It's very simple. You stab me in the back, I'm going to tit for Taty and stab you in the back the next time. And this turned out to be the optimal strategy. People then realized there was a problem, a problem with the tit for tat strategy in the real world, which was the possibility of there being a mistake. When I was a kid, there was one of these end of the world novels called Fail Safe. And in it, this was one of the big bad Soviet Union still existed in the United States and we all had nuclear weapons pointing at each other. And the starting point for the book is one day there's a glitch in the system, and one American fighter pilot up in his plane gets a message saying we're being attacked, go drop an atomic bomb on Moscow. And it turned out because there's a glitch of some sort of bug in there where it really hadn't been sent, he was mistakenly told that there was an attack happening and he should go take out Moscow. And what the rest of the book is about is, whoa, Moscow's saying, hey, we just spotted one of your planes about to come into the airspace. And the Americans saying, no, we didn't send a message. And then figuring out, oh, my God, we sent a message by mistake and trying to talk the pilot out of it. And the pilot saying, oh, I know by now you're Russians pretending to be Americans. I know it. And it's going to be a disaster. And the Soviet Union is saying, if you bomb us, we're going to release all of our weapons and it's going to be the end of the world. And, and they found a solution. They found a solution which was at the end on the very last page, what was worked out as a compromise. And this was early 60s. So this was like Kennedy and Khrushchev as the characters in it. They worked out a solution, which was OK, you've taken out Moscow because eventually he does get through and drops a bomb on Moscow and we're allowed to take out New York. Which scared the crap out of me because I was like 8 years old living in New York then and I lived in Brooklyn, which had the highest population. So of course I knew they would drop the bomb right on our roof in Brooklyn is terrified. But this was a solution for when there's accidentally a mistake being made, When you believe you have cooperated and the other side, because of the signal error, believes that in fact you cheated, you get one shot back on the final page. An American pilot who volunteered to do it drops an atomic bomb in New York because this is the way out of the signal error problem. OK, so let's run wild here with these little magnetic disks. Cooperate in the first round and there's a signal error. In the second round, A cooperated, but somehow there's a glitch and you get a message that they stabbed you in the back. There was imperfect communication. So you're B, and suddenly from out of nowhere, this son of a bitch stabs you in the back. And you say, Oh, yeah, well, next round I'm going to stab you back. Meanwhile, there's A who believes they've been doing nothing but cooperating, and they say, Oh yeah, you need to do that, How about this now? And they stab them in the back again and say, oh, tough guy, yeah, how about this? Now back we go. And what you see is you will get seesawing for all of the rest of time of cooperation interspersed with cheating. And what that does is with each one of these, they're building up lots of negative points. This is a disaster. This is a disaster if you get caught in one of these back and forths that was triggered by a mistake, a signal error. So what the game theorist had to figure out then was as long as your reality includes the possibility of a signal error, tit for tat is vulnerable because tit for tat will then get caught in this seesawing thing for the rest of the time and everybody loses. So they soon programmed Axelrod and Hamilton, now the possibility of signal errors into there, and everybody went through and submitted their strategies and out came the optimal strategy with that, which wound up being called forgiving tit for tat. What's that about? It opens the possibility that the other person didn't cheat, that in fact there was a signal error and you should ignore them having seemingly cheated. Maybe they didn't really so they cheat and in forgiving tit for tat you say Oh well maybe it was just a signal and you're screwed now because they've gotten away with once and you didn't hit them back again and it's a huge disadvantage. How do you do forgiving tit for tat and not get screwed? You only do it with someone who you have a long history of cooperation with. The longer the history, the more you can trust them, the more you have an established cooperative reciprocal relationship. And what you see is different versions of forgiving tit for tat. OK, if we've had three successful rounds of cooperation and there's maybe cheating, but maybe it's a signal error. We've gone 3 rounds, that's enough for me. I'm going to forgive them. Or what if instead you're criterion are if we've gone 11 zillion rounds of cooperation, I'm not a very trusting individual. Only then would I decide that maybe this was just a mistake and forgive them. And then seeing under different parameters when do you want to have how much of A requirement to forgive and seeing that depending on the histories and likelihood of signal errors and all of that, different ones worked optimally. But the main thing is that overall in this sort of world, forgiving tit for tat out competed tit for tat. Then you see, nonetheless, we've got a vulnerability here, which is if you know your opponent is using forgiving tit for tat, you cheat. And then because you've got this relationship set up, they forgive you. Yeah. And then you go a few more rounds with these nice salmon cold ones, a few more rounds, and then, yeah, you, oh, you know, I'm at a little cheat again. And then they forgive you again because they are too trusting. Forgiving tit for tat can be exploited by cynics who have a sense of how many routes you need to cooperate before they're going to trust you enough that they'll forgive you and say, ah, that's an anomaly. That must have been a signal error. Forgiving tit for tat is vulnerable to exploitation like this. What people then began to see was all sorts of other strategies. One of them was called contrite tit for tat. Another one was called Pavlov, which was a really mean exploitative one in terms of taking advantage of forgiveness on the other players part. All these different strategies and this turned into this whole field of seeing optimal ones. Under what circumstances do you take which strategy? And what people eventually figured out was that there was a cycle to it. So you start off with a population where everybody is a cheater and you have a mutation, metaphorical one, but we will see. Maybe not metaphorical. You have a mutation so that you adopt A tit for tat strategy and you start off by cooperating. What's going to happen to you? What a schmuck. Everybody else says, whoa, here's they don't want a fool being cooperative when So if there's one tit for tatter who suddenly says I'm going to make the world a better place and start off cooperating, they're going to be screwed. But if there's two tit for tatters and they could find each other green beers, if there's two of them, they're going to out compete all the cheaters. All the cheaters are going to have no choice to switch over to a tit for tat strategy. So hit for tat replaces all cheaters. Then if there's the possibility of signal error, forgiving tit for tat is going to outcompete tit for tat. At that point, if you have long enough of relationships, of cooperation, you drop this all together and your rule is always cooperate. Why does that wind up being advantageous then? Because you're having to put energy into surveillance for cheaters. Everybody's cooperating. Everybody's great now we're all cooperators. Forget it with having to have the alarm system there. You don't have to pay for that anymore. Don't bother even locking your front doors. Everybody's cooperating, and you select for everyone being cooperative. And what does that then do? That sets you up for one individual with a mutation as a cheater who pops up a wolf in sheep's clothing and can exploit everybody else. And thus there's selection for two cheaters and for, and before you know it, you're back to everybody being a cheater. You go through these cycles like that, where different circumstances bias you towards more or less rapid transitions. OK, so that's incredibly interesting. That was great. The game theorists were completely delighted at this point. And then suddenly the zoologists, I don't know, got stuck in an elevator with them or something and learned about game theory, and they said, whoa, this has to apply to the social behavior of animals. They have to use game theory strategies to optimize when they're cooperating and when they're cheating. And the zoologists went berserk at that point looking for examples of different prisoner dilemma strategies out there among social animals and seeing the same question. Is tit for tat the most optimal strategy? And they went and looked, and that's exactly what they saw. For example, vampire bats. Vampire bats go and drink blood. And that sounds really disgusting. But what they're drinking the blood for is they store it in a throat SAC and they fly back to their nest and they regurgitate it for their babies. Isn't motherhood wonderful? And that's gone. But vampire bats do something that's even more interesting in that they form cooperative colonies where everybody feeds each others babies. Wow, that's good. So if somebody hasn't gotten any food, never found a cow to suck blood out of, at least somebody else is going to be feeding her kids. It's a safety mechanism that's everybody is benefiting from cooperating. And what you see is initially they're all things like sisters or first cousins. It's initially driven by kin selection. But as it gets established, you'll even have animals in there who are not relatives. Whoa, reciprocal altruism. And what would cheating look like? And one of the bats. Oh, sorry guys, I couldn't found a find a cow tonight. OK. Everybody feeds her kids and then three days later she comes back and says, whoa, crummy luck. I couldn't find find a cow tonight either and all that. And what you see is she begins to be punished back for each time she does this. Everybody in the next round gives less food to her babies. We have a tit for tat strategy and if she's a good citizen at that point, next round everybody else goes back to cooperating. Mind you, this involves the totally bizarre, disgusting thing that you have to measure how much throw up each vampire bat is producing there. And there's a guy, this great zoologist Wilkinson, who has spent his entire career weighing bat through up. But that's exactly what it is. It's a reciprocal altruistic system with a mechanism for dealing punishers who start trying to be exploited if you punish them back. And tit for tat goes along. There one other version of this that people then spotted, which was in Stickleback Fish. Now Stickleback fish are not the smartest creatures out there, but nonetheless they've got some behaviors available and you can see one of their behaviors which is defending their territory. Here's what you do. You've got your stickleback fish in your tank there and you make it think it's territory is being invaded by another stickleback fish. You put up a mirror on the side of the tank there and these are not quite smart enough to say hey, that's me, that's me and I exist and I have my own self actualized desire. They say, whoa, some invasive fish that's coming to menacing and they defend their territory by bashing against the glass over and over and over, which thank God they're doing that because the other guy is attacking them as well. And they're doing this now. Give them a partner. Don't put a second fish in there. Put a second mirror in there that's perpendicular to the first fish. And now he's sitting here and he looks over and whoa, hey, bro, here's another guy there. We're now two of us, which is just great because there's now 2 invading males coming in. So you both go and you defend your territory and you have this cooperative relationship and what a great guy and we're both protecting our territory. Make the first fish think the second fish is cheating. How do you do that? Angle the mirror so the fish's reflection is set back, So along comes this invasive guy and you're defending it there. And you see this guy's going forward, but not as much as you. He's laying back. He's pretending to assist you, but he's not really. And you're sitting there saying that bastard, I can't believe it. I'm blistering my lips here on this glass. And Oh yeah, he's pretending to defend, but he's not really. And the next time you put the mirror up there, this fish doesn't bother attacking it. It tit for tats this guy back saying, OK, you left me one time holding the bag, having to defend this whole place. This time it's your problem. They even do that last couple of points. So people were so happy, so happy to discover that they were finding examples of tit for tat out in the natural world. And you can immediately plug that into evolution over the course of time. Stickleback fish and vampire bats who have the optimal strategy were winning in these competitions of cooperation and they left more copy of their genes and thus selecting for these kinds of strategies, selecting for it parentheses, assuming there's any genes related to it. You will then have evolved propensities towards cooperation, but also of cheating and defending against cheating. So this was going great until research being done on lions in the Serengeti. And this was being carried out by a guy named Crappo, Craig Packer, Craig Packer, University of Minnesota, and Anne Pusey as well, these two scientists. These two scientists were studying lions, and they would do the same playback technique. They would put a speaker with a recording of the sound of a big male lion, a stranger, a menacing stranger. And suddenly out of the Bush you hear this menacing lion and all the lions running on there. And that's great. And what he would notice was every now and then there would be this one guy who lags in the back, this one who didn't head in, one who seemed to be cheating, this scaredy cat who was pretending to go investigate this scary Bush. But this individual was actually staying back. So if tit for tat is the way the entire world works, everybody should be getting pissed off of this lion Lioness for not doing their share of it. And they should be mean to her afterward in some form or other. Less grooming, some such thing. And what was confusing was they didn't do anything. They didn't punish this cheating lioness at all. They were somehow being exploited by her. Oh my God, there goes all the logic of this. Until they realized, oh, she's not very good at that and she lags behind, but she's the best hunter in the group. That's where she earns her keep. So what we've just introduced is this great complexity here that there's multiple games going on at the same time and these social species, and there's crosstalk between when you cheat and when you cooperate, how they begin to make sense of really complicated social stuff you see in all sorts of species out there. So we will pick up with that.